Issuu on Google+

23 1 2012

Connecting with Cornell


from the Office of the Vice Provost for Research


06 t

30 t

60 t

88 t

94 t

100 t

Research In Progress

Research In Progress

Research In Progress

Undergraduate Research

Center Spectrum

Research Newly Funded

Seeing Is for Humans, and Computers

Using Light in Unusual Ways

Watching a Molecule Perform

News from the Research Centers

Selected Sponsored Research Awards

Tsuhan Chen

Michal Lipson

Scott C. Blanchard

Through the Lens of an Undergraduate Writer

More Research In Progress Through Page 87

106 t

118 t

124 t

128 t


Technology Transfer

Economic Development

Research In Focus

Seeing DNA and Beyond: The BRC

Laser Focused

A Lens on Opinions

Visualizing the Invisible





Seeing Is for Humans, and Computers


Peering into Social Networks


Movies, Pictures, and Visual Perception


To Discover a Dream Material


Using Light in Unusual Ways


How to Grow a Super Material or Troubleshoot a Classic One

Tsuhan Chen, Electrical and Computer Engineering


Jon M. Kleinberg, Computer Science

David J. Field and James E. Cutting, Psychology

J. C. Séamus Davis, Physics

Michal Lipson, Electrical and Computer Engineering UNDERGRADUATE RESEARCH


Through the Lens of an Undergraduate Writer


Agents for Seeing

David A. Muller, Applied and Engineering Physics


Landscape: An Image or an Encounter with Nature?



News from the Research Centers


CLASSE Looks Ahead



Maria Y. Park, Art


When a Poet Sees Lyrae van Clief-Stefanon, English


Single Molecules Trapped for Study Michelle D. Wang, Physics





Watching a Molecule Perform

u e

Scott C. Blanchard, Physiology and Biophysics, Weill Cornell Medical College



The Optical Biopsy



Selected Sponsored Research Awards OUTREACH



A Lens on Opinions

Douglas S. Scherr, Urology, Weill Cornell Medical College RESEARCH IN FOCUS


Optical Tools for Seeing How Disease Changes the Brain


Visualizing the Invisible

Chris B. Schaffer, Biomedical Engineering

On the Web


Flies and Humans: Looking into Genetic Variation in Populations Andrew G. Clark, Molecular Biology and Genetics


PAGE 106

Cover photo: Frank DiMeo Photos this page: Blanchard lab; BRC; Frank DiMeo; WCMC Urology

Robert Barker/CU

A Note from the Editor As we conclude the print publication of Connecting with Cornell with this issue, we present some of Cornell’s most dynamic and extraordinary faculty on the theme, ways of seeing. These researchers have said that being able to see—literally see minute action, whether inside cells in the human body or singular atoms in materials, or see for the first time how some social structures work—has not only uncovered much in today’s research and scholarship, but also offers much more for future advancements. Some develop instruments that enable researchers to see all kinds of processes in order to make groundbreaking discoveries. Some study how we perceive images with our eyes so that these human mechanisms can be copied for use in machines. Others visualize innovative ways of presenting both fresh and classic ideas. Still others use light in extraordinary ways. In these interviews and articles based on interviews with Cornell faculty, you will experience how researchers and scholars are helping us to understand our world better as they open up new ways of seeing, both literally and metaphorically.

What a finale for Connecting with Cornell! Although this is the last issue of Connecting with Cornell’s 20-year run, we will continue our coverage of Cornell’s world-class faculty. The Office of the Vice Provost for Research is committed to presenting Cornell research and its monumental problemsolving impact to the world. Be on the lookout for coverage in exciting new venues.

Ernestina Snead


Research in Progress


Seeing is for Humans, and Computers Tsuhan Chen


Images are represented by pixels, which are just numbers with values. We analyze these numbers and detect what’s going on in a picture. Computing the Visual We have all experienced the evolution of computer vision: photography enhancement, making the photos we take look better (making them sharper, improving the color, or removing granular noise); industry inspection, using cameras on industrial production lines to inspect parts as they are made; video surveillance, using cameras to analyze people going through airport security; even cameras that help drivers see better, like detecting a child in the driveway or an upcoming stop sign. And the movie Avatar has dazzled us. Our research has progressed steadily through these areas of computer vision technology.

Photos in Research in Progress: Frank DiMeo unless otherwise noted

Now, to this repertoire, we add research in the social sciences. We study pictures people take and share among themselves on social networks—analyzing the data they reveal. Our research is to understand the social relationships among people and how they develop. I use the term “visual computing” to represent everything we do that utilizes computing to process visual data: graphics, pattern recognition, computer vision, and more. What Your Facebook Photos Tell Us If you are an outdoor person, you are likely to upload pictures with trees, mountains,

rivers, or national park scenes in the background to Facebook. If you are an introverted person who spends time mostly in an office, you are more likely to post photos with computers or pictures on the wall in the background. By analyzing these kinds of images posted on Facebook, we know a lot about people. A company selling outdoor equipment, such as skis and bicycles, for example, already knows what you like by the pictures you post on Facebook. And the company will pay for the ad. In computer vision, also called pattern recognition, we use computers to recognize images or patterns in

Research in Progress

images that give us information, such as revealing what your interests are through the photos you post on your Facebook page. Sharing Images Online Proud parents often place pictures of their newborns and children on Facebook, Flickr, and other websites so that the grandparents can see them. We analyze pictures such as these and can connect the people within the family. A group of 10 users may register on Facebook independently, but through our analysis, we know that they are the



As engineers with our computers, we process, synthesize, and analyze numbers. Processing numbers, the first step, is easy. Generating numbers and making sense of them is the second step. The third step, making the numbers understand the real world, is the most challenging. And it’s the most fun part.

Photos Analyzing

Chen Lab

By analyzing the social relationship between people in a photo, we have been able to estimate their ages, identify their gender, and determine their identities.


At the highest level, our work is about the human, the family, and our visual interests. But under the hood, it’s about understanding images represented by numbers. same family because they share pictures among themselves.

analyzing numbers to see the patterns that the numbers represent.

By studying images, the people pictured in images, and people who post the images online, we gain an understanding of the social aspect of how people share images.

Photoshop is a good example for explaining image processing. If an image doesn’t have enough contrast, this means the pixels don’t vary enough. So we adjust the numbers to make the pixels vary more and improve the contrast.

How Engineers Do Social Science Collaborating with psychologists, we help analyze images that people draw—adult doodling, children’s drawings, and other sketches. A child three years of age draws something quite different from a 10-yearold. A three-year-old may draw a head on a stick figure. As the child develops, the human body develops further in the child’s drawing. By analyzing how children draw pictures, we can assess their development. The research can lead to early detection of developmental deficiencies in a child or changes in the emotional state of an adult. This is done in psychology all the time, but what we bring to the work are our algorithms, which help to analyze patients better. As the research continues to develop, much of our technology will be helpful to humankind through the social sciences. From Numbers to Pixels to Images to Information Our research has always been about looking at images, which are represented by what we call pixels. These pixels are nothing but numbers. For example, they can represent RGB color values. What we engineers have to do with pictures is what we do with the numbers that represent the pictures. We analyze numbers to detect what’s going on in pictures. Numbers can represent a human eye—the black circle of a certain radius with contrast against a white background that shows eyeball. A pattern represents a human nose and then below it, a human mouth. This is what tells us there’s a face in the image. This is engineering—

At the highest level, our work is about the human, the family, and our visual interests. But under the hood, it’s about understanding images represented by numbers. An Algorithm called iCosegmentation i = interactive, co = together, segmentation Let’s say you have a lot of pictures, and you need a specific object extracted from all of them. You don’t have time to take each picture, trace the boundary of the object, and crop it out. But if you give this task to our to algorithm, it will do it for you. All we need are the pictures and a few hand-drawn strokes on the object and the background you want cropped out. The algorithm will take care of the rest. It picks up the cue from you and propagates it to all the pictures. We created this algorithm and named it iCosegmentation. The task is segmentation. “Co” means we’re able to do multiple images at the same time. “I” is for the interaction you’ve had with the algorithm. An Actor in My Lab If an actor comes into my lab we can capture him with our multiple cameras from different perspectives and angles and render the images in the future, however they are needed. When a movie director requests the actor in certain poses, doing a particular thing, we can provide the desired shots. Movie directors find this very useful. We call this computer graphics—CG in the movie industry. We process the images to create special effects like the ones you

Chen Lab

Jason Koski/CU

Why this Research? I’m an exceptionally visual person. I am exhilarated by the things I see. We’re given , see, and understand our environment for a reason. So, I’m energized by what we call , which encompasses everything we do. Whenever I present my research results, people feel connected because they can see the results. I can show it to them. I love this connection with my audience.

Research in Progress

Chen Lab With camera arrays, we’re able to capture an actor from different angles. Once we process the information we’ve collected, we can regenerate that actor in many different configurations—different camera angles, facial expressions, or body gestures. That’s how Avatar was made. All of the computer graphic actors in Avatar are not just computer generated, but also driven by human performers. Avatar is a milestone in our research community.

The advent of touch-screen devices has opened new avenues for human-computer interactions. In this project, we study one such interactive application to allow a user to segment or cut out an object from a collection of images. We have developed a user-friendly system on a touch screen, which enables a user to do so by scribbling on a few images. The user touches the screen to make scribbles, and the system segments out the marked object from the entire collection.

Representing images as collections of visual words enables automatic object detection. Considering the 2-D spatial layout of the visual words, we have developed an algorithm to detect objects of interest efficiently. As shown in the figure, corresponding visual words (the eclipses) between the training images and the test image, with the same spatial layout, allow us to detect the motorcycle in the test image.

Which painting is more likely to be chosen from a large gallery for home decoration? How well can a computer predict peoples’ aesthetic opinion of a painting? We have designed an image-processing scheme, based on machine learning, to explore the relationship between aesthetic perceptions of humans and the computational visual features extracted from paintings.

Chen Lab

see in TV commercials and in movies like The Matrix. Seeing Like Humans A three-year-old child can see ice cream in your hand and say, “I love that ice cream. I want that ice cream.” Our computers cannot do this yet. Computers are good at detecting human faces and objects that are already somewhat known, such as cars, vehicles, and flowers. They can analyze the numbers that compose the image, but they cannot see like humans yet. For computers to see like humans, we have to move from analyzing images as an array of numbers to not only analyzing the pixels represented by numbers, but also analyzing the content of the image. Envision working on numbers so that the computer can detect a human’s eyes, and then the face. That’s the first step. The second step is to observe the human as the human moves about. The computer

must now recognize the arms, legs, and see that, “Oh, the person is actually running.” We call this body tracking. The movie industry uses body tracking a lot— tracking the actor as he performs and then rendering these actions so that we can see them again. Now, the next step is to analyze objects in the world and how they interact with each other. Take for example, your computer screen and keyboard. They’re not really connected, but when you look at an office and see a computer monitor, you expect to see the keyboard someplace nearby. Humans do this. The computer is still learning. These are the steps that lead to computers seeing the world like humans and having objects interact with humans. An Algorithmic Eye Imagine an algorithm—an eye—that can analyze objects in your office and tell you it’s time to clean up. If the computer looks at your office every day and sees the same

objects in the same location, this means you’re very organized. If the computer takes a picture each day and sees objects constantly changing places, then you may be unorganized and messy. Computer algorithms can be trained to do this. Eventually they will evolve to this point, and computers will be able to see the world as humans do.


Get the free mobile app for your phone


Peering into Social Networks A Conversation with Jon M. Kleinberg COMPUTER SCIENCE

How does a person—or anything—become popular? How do groups arrive at consensus? How do opinions form? These are some of the questions at the interface of computer science and the social sciences that fascinate us. You help us see the invisible networks all around us—social, political, technological, scientific. But let’s talk about social networks.


Social networks are made up of fleeting, unrecorded interactions: two people talk, a piece of information passes from one person to another, they tell another friend, and gossip spreads through the network. This is something that happens, but not something you could directly observe until now. KLEINBERG:

For as long as human beings have lived together in groups, we have had social networks. What has changed is our ability to see them.

We can see these processes unfolding on the web and internet, through Facebook, Twitter, email, and cell phone communication. They create traces allowing us to study the processes at a level of detail that has never before been possible. Science always advances when we can take things that were once invisible and make them visible. The ability to discern microscopic objects, resolve extremely distant formations in space, and image new kinds of particles that escape our direct experience have all led to big breakthroughs in their respective fields. This is happening now with social processes and networks. The challenge is how to take this

kind of information, which is nonphysical, and find good ways of representing it. What does a social network look like? To understand what a social network looks like, I first collect the structure of the network. A network is built of nodes and links. The nodes are the people—the individuals to be connected—and the links are the relationships among them. Let’s say I want to look at certain data: the social network defined by Twitter, for example. The nodes are all the users on Twitter, and we can define the links in several ways. In Twitter you declare yourself to

Research in Progress

be a follower of somebody else—I follow your tweets. The links can represent these follower relationships. If I follow you, there’s a link from me to you. On Facebook, it’s more basic. The nodes are the people, and the links are the friendships. A natural way to draw a social network is to start with the link between the two people, which tells us something—here are two people with a friendship. I might start with two people from very different social circles.

of Ithaca. This is not true of networks, partly because networks are nonlocal. I can’t maintain a tight focus in a network. If I zoom into one person, one step out I see all their friends, one step from there, I see their friends of friends, and we’re now talking about tens of thousands of people all moving in different social circles. Even two steps out—friends of friends—bring in so much complexity that I can’t break off a little piece of it.

What are the underlying principles that cause a group to end up in consensus or in conflict? Why do we form new links in social networks or join new groups? Person One is at the center of a tight cluster. Person Two is part of a cluster of people who all tend to know each other, but she lives at the boundary of different worlds, because she has friendships with clusters that are quite distinct. This translates into different life experiences for these two people. Person One tends to have a very orderly social life, but he doesn’t get exposed to a lot of new information. Person Two, living at the edge of different communities, has a more chaotic social life, but it gives her access to things Person One doesn’t know. Often, by just looking at a network’s structure, we understand something about how the network fits together. But you’re usually mapping this on the computer, right? When we first begin to study a network, we can draw informative pictures of it. But we quickly discover that if we try to build this using a million people on Facebook or Twitter, it’s a complete spaghetti mess. Speaking of mapping, it’s intriguing to compare networks of social structures and maps of the physical world. If I have an extremely detailed high-resolution map of New York State, I can zoom into Ithaca, see the streets, and not have to worry about Poughkeepsie, Rochester, or Manhattan. I can maintain a tight focus on the details


Networks are inherently complex. All the pieces are wired together, and everything affects everything else at short ranges. This is the small-world phenomenon—the six degrees of separation. That everybody is so few steps away from everybody else means I can’t treat this like a map, where I can zoom into a small part. This is why it’s hard to see into networks. At this point, we move to mathematical and computational tools. We need a mathematical vocabulary to talk about the landscape of the network—a way of looking at clusters of people and networks. What are you looking forward to revealing about social networks? What would be incredible to see or discover? When this type of data began appearing online over the past 10 years, the opportunity to address questions that lie at the interface of computer science and the social sciences emerged. First, let me talk about computer science. Once upon a time, the main constraints on the systems we built were technological ones—processor speed, network latency, and memory capacity. These issues are all still as pressing as ever, but they’re now supplemented by new concerns—the problems of designing for millions of users. Suddenly we have to worry about trustworthiness or the reliability of

information, the reputation of users, the similarities among different users, the problem of recommending contents to users—the social feedback effects from popularity. These are social and economic effects that have an enormous impact on how the system functions. In the end, we’re trying to design a computing and information system. Any of the large systems that have been built in the past decade, like Facebook, Twitter, Wikipedia, YouTube, and before these, Google, succeeded or failed based on how social feedback effects were managed—whether controlled well or out of control and breaking down the system. It would be fascinating to enrich the design vocabulary for computing systems with social feedback effects and social design principles, which ultimately will be embedded in all designs. And the social science research opportunities? This massive data offers a rich view of social processes, but at a scale we have never seen before. The potential to address long-standing questions in sociology and other areas of the social sciences is compelling. When this data became available, a number of us, including Dan Huttenlocher, Computing and Information Science, Michael Macy, Sociology, and I, sat down to discuss how fundamental questions in sociology— where the obstacle had been a lack of data at large scales—could be addressed in new ways and from a different angle. It’s another case where science can move forward because of something previously invisible now made visible. How does something grow from obscurity to popularity? How do opinions form? How does a group arrive at a consensus opinion or end up polarizing into two different groups? What are the underlying principles that cause a group to end up in consensus or in conflict? Why do we form new links in social networks or join new groups? These are wide open questions that we suddenly have the opportunity to address at a very detailed level using this kind of data. It’s exciting.

This is studying human behavior. Right. It’s a topic so complicated that it benefits from as many perspectives as we can bring to bear on it. The arrival of these data sets has made it possible for computer and information scientists to contribute to ongoing research on human behavior. With online systems, we can observe large-scale aggregate human behavior—how it plays out over time—in a way that had been extremely hard to observe. So who’s watching whom in this new age of online communities? This is an extremely important issue—privacy and our expectations of it. What privacy do we have online? We need to make one level of distinction between things online that are public, even though you may not realize they are, and things that happen online that are not public—your emails, instant messages, or queries you type into a search engine. These are not disseminated publicly on the web, but companies have certainly collected this information. When people put things online, you can find them using Google. They are public. Yet people don’t think through all the consequences of being public. It’s complex territory. There was a time when only a very small fraction of the population had to worry about the effects of public utterances—for example, people who worked in the management level of companies and got training on how to make public statements. But now, all of us are making public utterances and presenting a brand name to the world. Very few of us are getting any kind of coaching or training on how to go about this. We need to think about this area. And those things that happen online that are not supposed to be public? They present even more privacy issues: if you use Gmail, Google has your email. If you type your queries into a search engine, the search engine has your queries. If you make a call on your cell phone, the company has the information about who you called. When we talk about data that’s genuinely private like this, the question of what’s going to happen as companies acquire

larger and larger stockpiles of this kind of information—what they can reasonably do with it and what the consequences will be— is uncharted territory. Is privacy an area of research for you? It’s a question I think about. Studying massive data that contain traces of people’s behavior, I must think about the privacy implications. We can look at public data designed to publicly reveal one set of characteristics, for instance, but observe a

totally different set of attributes. In a paper we recently finished, we show that when people post photos online (photos have a name for the picture and when and where they were taken), we can figure out who they know—and not because of the people in the picture. If you take a picture of Rockefeller Center, for example, this reveals that you were standing in Rockefeller Center at a particular point in time, like 1:00 p.m. on December 6, 2010. If we look at the massive database


Why this Research? This field of research combines several long-standing interests. One is building mathematical models of situations in the real world. Another is applying this perspective to big, complicated problems. The third is about how computer science—which is about computers, technology, and the mathematical process of computation—has suddenly turned out to be a profoundly human subject, as well. The web emerged in the early 1990s, as I was beginning graduate school. That created a radical transformation in the field I was entering. Somehow the web—one of the most complex artifacts ever created by human beings—can only be fully understood if we think about the human components of it. It’s a profoundly human artifact to which we’ve all contributed. It reflects all the messiness of human life. It’s not clean and easy to search or to categorize. It has all of the internal contradictions and complexities we associate with the life of human society. And we can actually approach it computationally—it is an amazing blend of the computational and the human. It is thrilling to be able to study both at once. 16

Research in Progress

A Fascinating Challenge

In the space of 20 years, we developed resources where we can go online and have all of our questions answered immediately. This has clearly had a profound and positive impact on the world. of online photos, which includes Facebook, Flickr, and numerous other sites that host photos, it shows you and all the other people who were there at 1:00 p.m. in the same place. You’re standing in a crowd of people, all taking pictures. It’s not that you know all the people in the crowd. But what if you and one of those people uploaded photos from Chicago on the same day three months before? And what if you and that same person also uploaded photos from San Francisco on the same day eight months before that? After three or four times, we have increasingly high confidence that you know each other. We can now go and check social network information and discover that a social network link exists between the two of you. We don’t often think that as we post one kind of information, we reveal another kind of information, which we may not want others to know.

What data do you use in your research? We’ve typically used information that is online and public. It’s information that we download from the web just like everyone else. At the same time, we work with the institutional review board at Cornell, which reviews this type of research, because even though it’s public and online, researchers have to be careful about any consequences that can flow from using the data. We certainly are very careful about the research, and we work with people to make sure we look at this data in a reasonable light.


We might have hoped that once all this information was online, where we could read newspapers from around the world and see how different cultures and nations think about things, we might have increased our level of mutual understanding for other cultures. You could argue that this has not happened as much as we might have hoped. We know more about other cultures, but we don’t necessarily understand them any better. We don’t necessarily appreciate or sympathize with other perspectives any better. A captivating challenge is to think about how we can create tools that would bring about better levels of understanding and appreciation for perspectives different from our own. On a superficial level, it sounds impossible. But we’re the ones creating the online experiences for people. And we can design them one way or another way. We haven’t yet found a way to design online experiences that reduces conflict and animosity. But there’s a big design space out there. And that’s a huge opportunity for creators of social media.


Movies, Pictures, and Visual Perception A Conversation with David J. Field and James E. Cutting PSYCHOLOGY

Movies and pictures help us understand how the human visual system is organized. Together, you run a perception lab in the psychology department, but you approach your individual research a bit differently. David, you use mathematical modeling as a tool for looking at pictures.


F I E L D : There is an implicit belief that the human visual system is general purpose: that it can recognize and efficiently deal with any image. To understand how our visual system works and why it works, I have argued that we need to understand the statistics of the world we live in and how the visual system processes those particular statistics.

With modeling, I can simulate thousands of neurons and investigate how the population responds to a complex natural scene. I have found that this approach allows us to understand why visual neurons have their unique response properties. I also do perceptual experiments with human observers. I believe this combination of theory and experiment provides important insights into visual processing. TO CUTTING:

But you look at movies.

For the last four years I’ve been studying Hollywood movies. I claim that film structures teach us about the mind. CUTTING:

We have not evolved to watch movies; instead, movies have evolved to match our perceptual and cognitive systems. They have changed over time to match what we understand best and how we like stories organized. I study the temporal structure of the movies, how they change in time. I measure shot lengths, brightness, color, and motion and look at transitions between shots. I can make good inferences about the evolution of movies by tracking such changes and finding out how stories are told. In contemporary movies, for example, about 99 percent of all transitions between shots are cuts.

Research in Progress

James Cutting

Level 1: Urban Rain

Level 2: Hotel Interior

Level 3: Winter Fortress/Hospital

“Inception (2010) is a complex film. One of the ways that the viewer can keep track of the interleaved story is that the predominant color of each dream level is partly distinct. The upper images show averages of every frame at a given level, and the lower ones show an example frame.” –James Cutting Theory suggests these abrupt changes should be confusing, but they are not. We would like to understand why not. Part of the answer may concern a kind of blindness to quick change, and part probably has to do with it quickening the pace of the movie, allowing our mind to run faster. It’s fun to work with David because of his work on the spatial relationships in pictures. He has a long history of studying the composition of single images. As we look at organized structure over time and space, we can combine our expertise for both studying film and exploring what people might see during a walk across campus. What does your research add to the study of visual perception? F I E L D : We want to understand why the visual system processes the visual world as it does. Our work has focused on the structure of natural scenes. Research with static images has enabled many important discoveries. Now, however, we have the tools to study the different ways in which visual processes unfold over time and how the visual system sorts out both spatial and temporal stimuli.


Our visual system has evolved and adapted to work quite well in the world in which we live. A full model of visual coding, extending from the retina to object recognition, has not yet been developed. We’re looking at pieces of this—why particular neurons respond as they do and what their coding strategy might be. How are we able to recognize the patterns and objects that surround us? These neurons have a code that we haven’t deciphered yet. If we can decipher it, there would be all sorts of applications. But right now, we’re just interested in the code. C U T T I N G : The type of modeling that David does, where he’s analyzing an image in space, is similar to some of my analyses in time. The bet is that the regularities we find in the visual system for both will be the regularities that are furnished to us in the real world. By studying complex images in the real world and in movies, we can better understand the constraints on those regularities.

Might a computer scientist take your work and figure out how to make machines with better vision? Or might a movie director take it to learn how to improve movies?

Level 4: Urban Limbo

The New York Times and other press leapt on the idea that we were discovering how to make a better movie or how to make a movie earn more money. That’s not it. CUTTING:

But we may have tapped into how movies grab and hold our attention better. Over the duration of a movie, there is a certain set of fluctuations—a set of patterns that occur—and these seem to be related to the fluctuations that occur every day in our minds. The fluctuations in movies— shot-duration patterns or how often new shots occur—often mimic how we allocate attention over long periods of time. But making a movie that is good at grabbing our attention doesn’t guarantee a better movie. You’ve probably had this experience with a late-night movie on TV. You’re not ready to go to bed, so you turn on a movie that you would never have paid money to see. You sit down to watch it for a bit, and a half hour later, you’re sucked in. You ask yourself, “Why am I watching this?” It isn’t a good movie, but it has captured your attention, grabbing hold in a way that is very hard to relinquish. This may be the phenomenon we have tapped into. We may have discovered a pattern in movies that relates to how people pay attention. Are other future applications apparent for your work? F I E L D : No artificial visual system right now can compete with the human visual system. Artificial systems might have better resolution: we have great telescopes that can see farther and microscopes that can see smaller. But in recognition, our visual system is magnificent. We don’t know why.

Why this Research? James Cutting I have always been interested in perception, culture, and cultural artifacts. I wrote a book on Impressionist painting. I go to movies and art museums, because I love them. We humans create these things—films and paintings. We create them for particular purposes, and we try to structure them in particular ways. I’m fascinated by why we make these things and what we will discover about their structures.

David Field The visual system is the most well-studied system of all our senses. We know a lot about what individual neurons do, but we don’t have a general theory of why they behave as they do. The underlying code allows us to see a stable world, and it allows us recognize objects and process scenes, despite what appears to be a lot of very noisy neural responses. Trying to discover the underlying reason why these neurons act as they do is intriguing. It’s like breaking a code, and this fascinates me.

Making Sense of Visual Perception Research in the Field and Cutting perception lab covers a lot of ground:

How does the visual system respond to spatial-temporal information?

Why does our visual system prefer a particular class of images? Why do we select and place certain images on the web?

How do we perceive color?

Why do we have a remarkable memory for certain preferred images? What is characteristic about these images?

Starting several weeks before birth, a baby’s eyes generate waves of activity that spread across the retinae. These retinal waves are important to visual development, but it is not clear why. Do these waves help train the visual system, preparing it for the world the infant will encounter?

How do we make sense of abrupt, instantaneous changes in films? What is happening in the visual system during cuts?

How much information is in an image? How much can an image be compressed without loss? How much information is contributed by the different properties of the image, such as the contrast?

What allows higher-level neurons to be both highly selective and invariant? Can we provide a complete model of the processing required toachieve this selectivity?

How do we integrate contours? How do the neurons responding to each piece of the contour combine their activity to produce a coherent whole?

What is the structure of children’s films?

Research in Progress

David Field

“My work focuses on the statistical structure of natural scenes and how the visual system takes advantage of this structure. Images like the one above show many forms of statistical predictability. The physics of how our world reflects light produces images that are far from random. For example, neighboring pixels are correlated, the orientations of neighboring edges are correlated, and the textures of neighboring surfaces are correlated. My lab has shown that we can make sense of neurons in the visual pathway by understanding this structure.” –David Field

We’re still trying to understand how the neurons in the visual pathway allow us to be so good. If we can understand that, then a major application for this work would be to build a visual system that works as well as a human visual system. That’s what engineers and computer scientists are trying to do. F I E L D : They use much of our work to build algorithms: face recognition algorithms, for example. Many of these algorithms use the same early steps as the human visual systems—finding edges and contours and integrating them to represent objects. A good algorithm will almost certainly mimic properties of the human visual system, but even the best algorithms cannot yet replicate it. This is a good example, because in low-resolution contexts, such as seeing a person at a great distance, the face is almost not there, but we know it’s a human form. It is as if higher regions of the brain say, “This is a person,” and they then constrain the inputs to help pick out the features that make a face. Our visual system is not passive. We are knowledgeable perceivers.

What would be most exciting to discover in your area of research? F I E L D : A few years ago, I started a more general theory of why neurons are tuned the way they are. This required that I collaborate with researchers who have more extensive mathematical skills. The project is to understand the very complex properties of neurons and to simulate their behavior.

Neurons higher up the visual pathway typically become more selective. They may respond particularly well to a single object class or even a particular individual. One particular neuron, for example, was found to respond when an image of Halle Berry was presented to a patient, but did not respond to other faces or actresses.


These highly selective neurons, however, are also more invariant. The neuron would respond to Halle Berry independent of her particular pose, lighting, or clothing.

F I E L D : Researchers found other neurons that were selective to Jennifer Aniston. This tradeoff between selectivity and invariance is very interesting, and how this is achieved is a big issue in visual science. If we could fully understand how this works, it would be significant for those building artificial visual systems. C U T T I N G : I’m completely absorbed in the structure of films. We have analyzed 160 films from 1935 to 2010. We’ve had a large number of undergraduates helping with this. Some results are that movies now have more motion, they are darker, and they tend to have more scenes than in the studio era up to about 1960.


And it didn’t matter whether she was dressed as Cat Woman or was uncostumed. CUTTING:


To Discover a Dream Material J. C. Séamus Davis


We watch electrons—a billion times smaller than atoms—in complicated materials to enable the unimaginable materials of the future. Seeing Electron Waves To see miniscule subatomic quantummechanical waves shooting around in a piece of material is beautiful. It’s like looking at Cayuga Lake on a breezy, blustery day: you see waves of large amplitude, lots of short waves, and long waves all interfering with each other, making a very complicated visual pattern. That’s what I see—the same kind of wave patterns, but now quantum mechanically—when I watch electrons in materials. I had to invent a scheme to be able to see this. To develop the unimaginable materials of the future, we need to see how electrons move in complicated materials.

Complicated 21st -Century Materials We have known for 50 or 60 years how to design instruments and materials, like silicon, gold, aluminum, and platinum, for use in our technology. But the new electronic materials being developed in the 21st century are immensely more complicated than a familiar material like silicon or gold. It has been extremely challenging to understand the properties of new materials—how and why they work. One of the reasons is because scientists could not see directly what electrons are doing in these materials. But as soon as we developed a technique to visualize the action of the

electrons, we saw that it’s extremely complicated. My lab develops tools, approaches, and instrumentation to tackle the complications—to visualize what’s called complex electronic structure, which describes how the electrons work in complicated 21st -century materials. From Seeing Atoms to Seeing Electrons The technique my lab developed is called spectroscopic imaging scanning tunneling microscopy. It works like an atomic-scale gramophone. If you think back to the way an old phonograph LP works, a needle moves

Research in Progress

Davis Lab

Waves in the ocean look different as they splash around a lighthouse. Like a lighthouse in the ocean, this is a lighthouse of zinc surrounded by waves.

Shows a way of visualizing the lengths of different electron waves in the material

around on the surface—it’s a moving plastic disk with wiggles on the grooves, and a stationary needle conforming to the wiggles. This mechanical motion sends a signal to the audio amplifier, which sends a signal to your ear.

a fabulous device, but we can’t see electrons with it.

Now shrink that apparatus down by a factor of a million. The tip of the needle is one atom, and the surface has a small number of atoms in it. You move the needle around and measure electrons jumping from the needle to the surface. By measuring this electrical current, we can take an image of where the atoms are.

My challenge was to go beyond that state of the art to a new level, where instead of seeing the atoms, we could see the electron waves. I took the same approach of atomically sharp needles scanned over a surface, but I measured a different physical property, which concerns electrons and not atoms. We created an instrument that lets us see electron waves as they zoom through crystals. It was difficult! It took about 15 years and about 100 man-years to develop the instrument to its present capabilities.

This apparatus, invented in the mid-1980s by two Nobel Prize–winning physicists, is called a scanning tunneling microscope, an STM—the first device for seeing atoms. It’s

The Search for New Super Materials In your home, you have ferromagnets— refrigerator magnets—a quantum-mechanical material. It’s magnetic because the electrons


zoom around in circles and never stop. The magnetism does not decay. It survives forever. Other possible quantum-mechanical materials may exist, which we could use in our technologies. But we’re searching for them. We know they can exist theoretically, but we have not developed them to the point where we can use them in our phones and iPads. One of them is called a superconductor. It’s a material through which electrical power can pass without any dissipation. Your laptop would never get hot if it’s a superconducting laptop, because it would not use any energy. It would also work a thousand times faster. Superconductors need very low temperatures before they can work. We need room-temperature superconductors,

Why this Research? The fundamental issues in this field are some of the most profound in physics. Although we talk in terms of making new materials, iPads, saving energy, and so on, for a physicist, there are deeply profound questions posed by the existence of complex materials. It’s fascinating as a professional to have an opportunity to learn about and attempt to solve some of these problems. I want to mine down and get to the lowest, most elementary, most elegant level of explanation of what’s going on. I’ve had marvelous opportunities and very good luck. I also have wonderful colleagues at Cornell with whom to collaborate in this area of research. I’m delighted to have a career as a physicist. I wouldn’t change it for anything. 28

The Spectroscopic Imaging STM A lot of new software, hardware, electronics, and mechanical devices is welded together into a large, complicated instrument several meters high and several meters square. Deep in the heart of it is a little needle with one atom on the end, scanning over the surface looking at the atoms. By measuring the differential tunneling conductant, we watch the electrons and see how they move through the materials. It takes all of this and much more to understand complicated materials and make better materials based on that understanding. I invented the technique spectroscopic imaging scanning tunneling microscopy—which gives us the ability to see quantum waves in complicated materials. Only a few machines like it exist in the world and the majority of them are at Cornell.

but we do not yet know how to develop them. For the ones we know about, we have not figured out why they are so complicated. Figuring Out the Complication One of the things we discovered here at Cornell in the last few years about the complication of new materials is this: when we look at them directly at the atomic scale, electrons are not simply moving around through a piece of silicon. They are in a complex, self-arranged relationship in the material. Neither the theory nor the experiment exists to discover or develop new materials under these circumstances. Fundamentallevel work has to be done first to understand the basics. My lab is trying to figure out how the electrons cooperate together, so that they can travel through a material with no loss of energy. We hope this knowledge will help lead to the discovery of room-temperature superconductors. What’s Super about Room-Temperature Superconductors? We have longed dreamed of having superconducting computers. With room-temperature superconductors, you would not have to recharge your iPad battery, and its efficiency would be much better. Now, think about Google’s server farms, where each server farm with millions of computers is bigger than a football field.

These server farms have one of the most rapidly increasing rates of energy usage of anything in the world, because every classical computer uses lots and lots of energy. Superconducting computers would not use this energy. Instead of using megawatts to run a server farm, we could use a kilowatt and get far more computing power. Another example is the new photovoltaic solar cell arrays, which will be built in southern California, Arizona, New Mexico, and other places. Power needed in New York City is generated in the desert. We cannot build a high-voltage power line across the whole country from every wind farm and photovoltaic installation, generating gigawatts of power from sunlight. We would need an efficient and unobtrusive new way to send power from remote and ecologically appropriate power generation sites to where the power is needed—East and West Coast cities.

One way to do this would be to use superconducting power transmission. We could send enormously more power through a nondissipative superconducting cable than through a regular copper wire—and at zero voltage. With cheap room-temperature superconductors, we could revolutionize power generation and transmission. It would change the world.



Using Light in Unusual Ways A Conversation with Michal Lipson ELECTRICAL AND COMPUTER ENGINEERING

Little devices, including an all-optical switch, all-optical modulators, even cloaking devices—this is what the ability to manipulate light can yield. And Lipson has invented many little “light” devices for use in the microelectronics industry. SNEAD: You actually manipulate light to make

devices. How do you force light to make things? L I P S O N : Light usually likes to propagate in straight lines. So how do we tell light, “No, no, don’t go straight! Turn left. Turn right.” How do we distribute light? This is a very hard thing to do, especially at extremely small distances. Since light likes to go in a straight line, we can put it in a fiber, and then bend the fiber so very little that light doesn’t notice it’s bending.

Now, how do we make little devices? A lot of what I do is for computer applications. That’s chip scale. We use a special combination of two very different materials—oxide

(basically glass) and silicon. These are the same materials used for microelectronics. It happens that this combination is fantastic for bending light. Light always wants to travel in a straight line, unless we use this combination of materials to force it to bend. We etch down the silicon to form miniaturized fibers on the chip. About 10 thousand times smaller, the silicon fibers are now called the waveguide, and light travels through them. We actually make tiny tubes of silicon through which light travels, and we bend the silicon so that light stays in the silicon and travels around.

Once you know you can bend light using these special materials, what’s next? Our goal is to make photonics or optics for applications in microelectronics, where they’re really needed. The microelectronics industry today is limited by heat dissipation. The industry can’t really increase the performance of next-generation computers because computers are getting hotter. But if we use light to propagate information, we can transmit a lot of information very fast with very low power dissipation. That’s our goal. Is that what you mean by all-optical information systems?

Research in Progress

Yes, all-optical switches, all-optical modulators—all the little devices that compose the entire link of sending, distributing, switching, modulating, and detecting information. Everything is controlled by light. You’ve made many devices by controlling light, making it do what you want. We’re getting very close to where we can build systems that are made of all our devices, exhibiting formats that no one ever dreamed could be done with standard technologies. The amount of information we could send using a system made of these elements would be unbelievable. We are talking about supercomputers on a little chip, for example. Right now, what would be your dream device? Now that we know how to control light, we are looking at other areas outside of microelectronics where our research can be



Several areas in which I’m interested are a bit more exotic than previous ones in the field. My father, Reuven Opher, also a physicist—has been a strong influence on my career. He always encouraged both me and my twin sister, a professor in space physics at Boston University, to think big and look for more risky, high-impact ideas. With the MacArthur award, I’m now putting more emphasis on such projects that lack standard funding. One example is the ability to transfer heat or coolness over relatively large distances. Today heat is transferred mostly though conductivity, which requires one object to touch another. My group developed a new technique for transferring heat just through special waves called phonon polaritons, which can propagate largely in a vacuum and do not require an actual medium to propagate efficiently.

We can actually take pictures of some of the larger devices with high-resolution microscopes that are sensitive to the infrared light that’s being guided. You can really see light going by. applied. We can control light using similar materials for a completely different application. For example, using light we have demonstrated cloaking—the Harry Potter phenomenon. The cloaking device is used for hiding defects in objects. In the fabrication world, an interesting application is for masking. Say you make a mistake, and you don’t want to refabricate an entire thing. That’s okay—we can put something on top of the defect and mask it out. Based on our work, a group at MIT demonstrated cloaking using the same principle, but now on a larger scale. They can actually hide something big—on the centimeter scale. Now we know that bigger things can be cloaked. Of the many devices and techniques you’ve invented, which one excited you most? We have about 20 patents based just on the way we bend or use light. One of the most popular devices—it got a lot of attention and has had a big impact—is the modulator, which translates electrical signals to optical signals. We were the first to demonstrate a very small, very low power modulator. This device opened a new field, silicon photonics, which was rather exotic when it began. Silicon was not used as an optical material. It was only used as a microelectronic material. But once we demonstrated it, we got lots of followers. All major industries and universities now have research in this area. How do you show us, the nonscientists, what you see? How do you create images? We can actually take pictures of some of the larger devices with high-resolution


microscopes that are sensitive to the infrared light that’s being guided. You can really see light going by. You also developed a laser with your husband, Alex Gaeta, professor of applied and engineering physics. It’s not exactly a laser—it’s an oscillator. It emits like a laser. The idea is that we send light into the chip, and light travels around and around. We make a wave and a loop, and it gets amplified. As a result, we get a whole series of colors all with the properties of a laser, meaning they are very directional and are collimated. These oscillators could be used for computer applications, to provide the light that is needed to send the large amounts of information across the chip. For optics, it’s critical to have many colors of light. That’s the beauty of light—we can transmit information using not just one color, but many colors, all propagating together, which gives us much more bandwidth, much more information. I note that one application of your work is sensing. Are you looking at making sensing devices? That’s another application of manipulating light on a chip. We have a project with Antje Baeumner in Biological and Environmental Engineering, and we work on sensing analytes using optics. It’s a spectrometer on a chip that could detect the spectrum using very little specimen, instead of putting on a specimen and waiting for a few days. Our sensor will be so sensitive that it will work with just a few analytes. This is an important, very fast measurement for detecting diseases early.

Lipson Lab

Jason Koski/CU

Lipson Lab

Why this Research? The beauty of this research is its great combination of fundamental physics with very clear applications. I like that a lot. And I can make a difference using interesting physics—invent a completely different technology based on basic principals of physics and apply them directly to applications. I believe that, as researchers, we have the chance to open up new areas to explore and show the possibilities. And we’re definitely going in this direction, making sure that Cornell is always a leader.


Research in Progress

Lipson Lab

The Force of Light We have recently demonstrated that light can have sufficient force to move and control mechanical oscillators—the type that are embedded in our cell phones, for example. We showed that this ability can enable synchronization between such mechanical oscillators. Because light can propagate at long distances this means that, in the future, a cell phone with such a mechanical oscillator could force another phone in a completely different city hundreds of miles away to tune to the same channel, as long as they were linked in some way through light.

And you’re starting a company?

You do so much, and you’re a mother, as well.

Yes, with Alex Gaeta. The name of the company is Picoluz. It’s pico, meaning little, and luz for light. I’m Brazilian, and Gaeta is Spanish, so altogether it’s “luz.” The company is based on little photonic devices for nonlinear applications—for example, amplifiers and light meters—all based on nonlinear properties of silicon photonics. The market is instrumentation for making measurements in the lab.

When I got the MacArthur Fellow award, it was the first time my boys really connected to my scientific world. They came home from school saying, “Ima,”—they call me mother in Hebrew—“Everybody’s asking us about you!” My youngest son came home with a congratulatory note from his teachers. It felt wonderful. Among scientists, there are not many female scientists and definitely not many mother scientists. To get this endorsement from my kids is very special. They’re ages 7 and 15, and both want to be scientists, although I’m not sure they really have a choice [lots of laughter].



How To Grow a Super Material or Troubleshoot a Classic One David A. Muller


Being able to see each atom of a material has allowed us to troubleshoot, discover, and create materials of all kinds. The SuperSTEM and Us Material failures can cause devastating accidents or take down entire systems because a few atoms of the material are in the wrong place. Figuring out what caused a failure or problem and rectifying it is the suitable solution. The answer may also be to create a new, better material. But first, we have to see the atoms that make up a material. Using our electron microscopes that we perfected here at Cornell, I can see a material’s makeup, atom by atom. I can work out what each atom is and what each atom is doing. Developing techniques for our instruments and using them in materials

research has led my lab to substantive discoveries as we create, characterize, and control materials at this very, very small length scale. To give you an idea of the scale at which we’re working, imagine that a wafer of computer chips is the size of the United States; a single transistor on a single chip is the size of a car parked somewhere in the United States, and that car has a pin in it. We can find that pin somewhere in a car parked somewhere in the United States, and tell you the color of the pin. This is the scale of magnification we have

with our electron microscopes—the Nion SuperSTEM (scanning transmission electron microscope) and the F-20. So the challenge is, if we are looking at trillions and trillions of atoms in a piece of material, where do we start? We start with something you can handle with your hands, perhaps 30 centimeters across, and chop out a piece of a few microns. We use micromanipulators and focused ion beams to carve out little sections of the material that are just a few atoms thick. We transfer that to our electron microscopes and look at the individual atoms in it.

An Electron Microscope Our two electron microscopes do slightly different processes. We need both in order to do our work. The Nion SuperSTEM takes color pictures, where every color is not just a different atomic species but also a different electronic state of each species. It gives us color maps. The SuperSTEM is a great tool for discovery. Cornell has the first of this new generation of electron microscopes designed for chemical analysis and imaging at the atomic scale. Only four exist in the world, but others will follow.

When we zoom in very close with our electron microscopes, we see that graphene is a layer of common carbon atoms whose structure looks like chicken wire—lots of little hexagons, but it’s only one atom thick. It consists of many patches that have grown together, stitched at the boundaries like pieces of a quilt. Each piece of material is a little scrap that cannot grow big enough by itself, but the pieces stitch themselves together reasonably strongly. Each patch in the quilt is a different grain with a different orientation. When we color it by orientation, it looks like a patchwork quilt.




What we see – the structure of materials, atom by atom What we use to see them – Aberrationcorrected and monochromated electron microscopes, used with multiple Cornelldeveloped techniques To what benefit – to see what has not been seen before, in order to make new and better materials, including those that do not exist in nature To see one atom at a time, we have perfected these machines and developed many techniques for them at Cornell.


If we have a collection of computer chips, and one transistor on one chip is not working properly, we would take our focused ion beam—which is like a little trench digger or excavator—and cut out just that one transistor. We pick it up with a microscopic needle and put it on a grid. Then put that in the microscope and shine the electron beam on it. The electron beam goes through the sample and forms an image on the screen underneath, and we can see every atom. Just a few atoms in the wrong place will stop a computer chip from working, cause a turbine blade to crack, or poison an industrial catalyst. It’s a bit like CSI forensics, except with single atom sensitivity. Growing a Super Material My students and I are working with Paul McEuen, Physics, and Jiwoong Park, Chemistry and Chemical Biology, and their groups to learn more about the properties of a new supermaterial, graphene.

Until we took these pictures, no one really knew exactly how these pieces join up or how big they are. The really cool thing about graphene is that this one atom–thick material can be grown into sheets that are almost a meter across. We have been able to image all the different patches to show that graphene, even with the patches, is still a good conductor. The first applications for graphene might be things like flatpanel displays for your cell phone or large screen TV and other similar devices. It would make an excellent electrode for solar cells. Discovering Soil Facts My lab is having fun looking at soils from the Amazon and other tropical regions with Johannes Lehmann, Crop and Soil Sciences. Adding biochar to these soils can greatly improve their fertility, which is surprisingly not very good—the nutrients just keep getting washed away. Learning how a soil is put together at the atomic scale is intriguing.

Why this Research? I was going to do electrical engineering and build faster computers. I realized that I could not take a bunch of chips and string them together in parallel, because things were too slow. I thought about how to make faster transistors and recognized I needed new materials to do that. And those new materials had to be designed at the atomic scale because that’s how small some parts of transistors are now. When you get down to things that small, you cannot look up the properties of materials in a table the way device simulation programs do. Materials take on new physical properties just by virtue of being small. Understanding the physics is a must. And it is imperative to have some way of knowing what is happening at that small scale. The most effective way to do this is with electron microscopy.

Research in Progress

Muller Lab

Graphene has lots of little patches that have grown together. Grains rotated in one direction are green, and grains rotated in another direction are red. The little black dots are the stitching at the grain boundaries. This image shows about 250,000 atoms across, thinner than a piece of paper.

Traditionally, scientists think of soil as big blobs of this and that—we know what each blob is, because we can look it up in a table. But when we look at soil with an optical microscope or any x-ray synchrotron, it’s messily intermixed and looks like a blur. It’s confusing.

Silica Gla ss

Muller Lab

In collaboration with the University of Ulm, we accidentally discovered a two-dimensional silica glass—just one structural unit thick—making it the world’s thinnest pane of glass. For the first time, we were able to see the location of all the atoms in an amorphous material and directly confirm a model first proposed 80 years ago.

When we look at it with our electron microscopes, we realize the soil is made up of nanoparticles that are just 5 to 10 nanometers across, and some of them are only 2 nanometers across. The nanoparticles have very well-defined shapes, structures, and chemistry. Now that we realize what the nanoparticles are, when we look at soils from Australia, we see that they are composed of the same building blocks as the soils in the Amazon, but the building blocks are so small that scientists missed them previously. At previous lower resolutions, what looked like iron, carbon, aluminum, and silicon stacked on top of each other, is actually in little separate pieces. The result is we may get a new way to understand, classify, and predict the properties and behavior of soils, so that we can improve their fertility. Troubleshooting GM’s Fuel Cell The last several years, we have worked with General Motors to help make a fuel cell vehicle more durable. A fuel cell vehicle can have a 300-mile range—you can drive it from Rochester, New York to New York City. My students have had fun driving the vehicle, powered by fuel cell, at high speeds on backcountry roads, as well as around the Cornell campus. It works just fine.

GM was how do the catalysts degrade at the atomic scale? We were able to look at the catalyst material with our electron microscopes at different stages of the fuel cell’s lifetime. From that observation, we identified two distinct degradation mechanisms. The debate over which of these mechanisms was the culprit had previously divided scientists into two camps, arguing for decades. No one really knew the answer. With the ability to look at this material directly at different stages of the degradation, we identified the actual mechanism by which the catalysts degrade. It is a combination of the two methods, where each by itself does a little bit of damage, but the two methods together greatly accelerate the damage. If you slow one mechanism, you’ll slow the other as well. Scientists in the field had not considered this. We helped GM build a better fuel cell. We figured out the problem, and they changed their process.


The problem is that the fuel cell does not last as long as it needs to in order to be an economically viable vehicle, because it contains very expensive catalysts with platinum. And the catalysts degrade over time as you use them. The big question for


Landscape: An Image or an Encounter with Nature? Maria Y. Park ART

As we go about living, in so many ways instead of looking and actually seeing, we’re just seeing the projections of things. We see not what we’re actually looking at, but what we think we want to see.

Do You Really Experience Nature? What if our notion of nature is specified through its representations and not through nature itself? Our perception of nature may not be about nature, but about the way other people want us to experience it—like a consumable object. For a long time I’ve been interested in how media prescribes the way we should understand places or experiences through the containment of an image. I want to look at that image, take it apart, and then add a “delay” to the way we experience it.

Counter Nature Recently I completed a three-part series titled Counter Nature, which presents encounters with nature as against its cultural construction. The series is also in response to nostalgic visual conventions of landscape and the packaged media images of nature that deactivate our looking and further render nature as a remote concept. Through a series of paintings, painted objects, and photo-transferred cubes, I explored the distance between the embodied experience of nature and its visual representation.

Photos in this article: Frank DiMeo, Maria Park, Alan Shaffer unless otherwise noted

The title, Counter Nature, is a double entendre referring to a site of transactions with nature, as well as contradictions of the natural. In the first two parts of the series, I worked with images of national parks, like Niagara Falls and Hanauma Bay in Hawaii, looking at vantage points outside of the directed encounters. Instead of the commanding falls, for example, I worked with less iconic scenes like the parking lot or the hillside next to the falls. The third part of Counter Nature focuses on the moment of happenstance created

Research in Progress

Counter Nature, a series in three parts: “Through an orchestration of a shallow, spaceless depth of field with an intensification of brilliant colors in brittle fragments, Park engenders heterotopic landscapes and public spaces that are vividly imaginative and undeniably real.” –Patricia C. Phillips Counter Nature 1

Counter Nature 2

installation view, Margaret Thatcher Projects, New York, New York (2010)

installation view, Toomey Tourell Fine Art, San Francisco, California (2010)

Counter Nature 3

installation view, Sabina Lee Gallery, Los Angeles, California (2011)

2.5-inch cubes, acrylic and transfer on acrylate (2011)

You’re seeing one image on the top of the cube. When you turn everything around, it prompts you to ask what you’re looking at. It’s a mirror image of the other image, and now you’re suddenly asking yourself which one is the mirror image of the image? It almost becomes a game.


Why this Art? Even though I grew up in a family of scientists, I became an artist. I was interested in finding the between things that are categorically very different, like an architectural space and a musical performance or a relationship between two people. This is the way things percolate for me. Things that are oblique or seemingly extraneous become integral to how I understand something else. Art allows me to and think in this way. The studio is also the one place where I can set my own pace. Nobody tells me how long I can spend on something or what its resolution should be.


Research in Progress

My most recent work is a commission by the Johns Hopkins Hospital in Baltimore. The project consists of large-scale paintings, which are site-specific installations for their new building, Sheikh Zayed Tower, opened to the public May 1, 2012, and designed by Spencer Finch, with concepts of the garden and the natural world.

Acrylic on Plexiglass (2011)

I continued exploring some of the ideas formed in Counter Nature 1, emphasizing the play between the organic/natural and the geometric/ designed with a measured balance between anticipated events (such as the play of light and shadow) and unanticipated events (such as the appearance of graphic stripes and foreign color) in the painting.

upon encountering a crowd prior to knowing the exact nature of the event. One day I was passing through Pittsford, New York, and saw a crowd looking at something. It was a moment when everybody was looking. I took the middle ground and the background out of the paintings in this series to focus on that moment and to play with the concept of making the foreground into its own set of spatial distances.



One of my key interests is how the world is obsessed with acceleration, especially in the many ways technology intervenes in our lives. The ultimate freedom is for individuals to set their own pace—you decide when something is going to be done or when something is going to be experienced. But there aren’t so many opportunities for this—an experience that can take you out of a streamlined way of life and you have a moment, or a “delay,” when you can really ask yourself questions. I want the moment when you see my work to create this space in time.


Between the Viewer and the Object The Counter Nature series includes reverse paintings behind transparent sheets of Plexiglas and polycarbonate, stenciled paintings on Plexiglas cubes, and phototransferred Plexiglas cubes. I explored geographer Carl Sauer’s notion that any kind of land, whether a national park or a person’s property, are cultural designations by wrapping an image around the cubes. The corners of the painted cubes start to dissolve as the picture plane seems to hover away from the object. Counter Nature is not about the image. It’s about the behavior of an image, both with other works and with the viewer. In some ways, the work resides somewhere between the viewer and the object. The work becomes most interesting to me when the interaction with it reaches the same kind of cadence as the embodied experience of nature. Life Informs Art Having young children, I’m rediscovering developmental toys and find blocks, or cubes, to be inherently perfect. A block can be a discrete object or a part of a whole; I can take a block away from a group, and it is not diminished in any way. This concept of discrete presence happens in art. Working on this scale and in this medium gave me a way to continue my work while I was with my daughter.

Sometimes the mere practicality of life informs the way I work. Mystery versus Transparency Sometimes artists like to maintain mystery about their process, but I like revealing or having a certain kind of transparency to the process. The potential that knowing doesn’t simplify the effect of the work is what makes this activity so engaging. I love that people can see the intended image, but they can also see that this is not all they’re supposed to observe. All the unpacking of a certain image through multiple ways of looking at it becomes the work. Authentic to My Experience Although I’m a visual artist, I don’t usually remember how things look. Instead, I remember how they make me feel. I work on a piece until it feels authentic to my experience. While working on a painting, a person wearing red or flowers in a landscape may suddenly appear. I did not see the flowers until I painted the hillside, and suddenly I see the flowers there.

` /faculty-profile.cfm?customel_ datapageid_7102=22564

Seoul National University

“Manifest Destiny” installation view, Seoul National University Museum of Art, Seoul, South Korea (2009)

When a Poet Sees A Conversation with Lyrae Van Clief-Stefanon ENGLISH

Do you see yourself when looking into the mirror or how someone has described you? How much of what you see is being reflected back from society? I envision that how a painter, writer, poet, or any creative artists see the world is reflected in their art. Is this true of your poetry?


VA N C L I E F - S T E F A N O N : All the strange things I come into contact with—get obsessed with or fascinated by—find ways of popping up in my poems, which is exciting. For example, once I spent time at the Johnson Museum looking at bowerbirds. Later, I talked with my sister on the internet, and as we looked at them, trying to figure out their strange individual aesthetic, we also wondered how they would end up in my poems.

What was your fascination? I’m fascinated by their sense of design and the things the males collect to attract a female, like little snippets of blue bottle caps, which they put around their nests. They’re using found objects for art, but it’s art with a purpose passed along in their DNA. If they don’t design a nest that the female likes, they’re not going to mate. How do you think the bowerbirds will end up in a poem? I never know how things will show up

initially. I’m an avid researcher. Sometimes I have to force myself to stop researching and sit down and write, because I’m constantly collecting information. I don’t know how it’s all going to weave together. Oftentimes, I’ll write a line, and the language shifts in different ways, like by sound. I can be writing about time. Get the word “hour” on the page, and suddenly, here comes that bowerbird, because I said “hour.” It’ll wind up in the poem in the exact space where it’s supposed to be, just because of the rhyme.

Research in Progress

I selected the poem “Urania’s Mirror” from your book, Open Interval, to discuss. What’s it about? In the 19th century, there were illustrated cards for teaching the constellations. The cards had holes punched out where the stars would be, so you could hold them up to the light and see a constellation. The boxed set was called “Urania’s Mirror.” Poets return often to the idea of Urania’s Mirror and the universe reflecting things. It’s the first poem that I wrote for Open Interval. I think in terms of collections, instead of individual poems. I want the poems to be in conversation with each other, within and across collections. I leave Black Swan, the previous book, enveloped in a lot of mythology. The last poem in Black Swan is “Helen.” The book begins with Leda and ends with Helen, her daughter. I cover the whole mythological story of Leda and the swan, Zeus coming to earth in the form of a swan, Helen, the war, and everything that results from these events. Helen looks at herself reflected over and over again. She looks both at herself and for herself, because she’s supposed to be the most beautiful woman in the world. I’m in conversation with a Rilke line: “Beauty is nothing but the beginning of terror.” Her beauty has engendered a lot of violence, so she tries to see herself, and she catches a glimpse of what she thinks is the real her. The last thing she says is, “Girl, you’re lookin’ right through me.” The poem, “Helen,” is about the notion of looking for oneself through the mirror of words and in literature and mythology. In Open Interval, I also wanted to be in conversation about not wanting to do the same thing over and over again. There’s a sense of frustration in the way the narrative structure works in the poem: “Here we go again.” So “Urania’s Mirror” starts with that: “Here before your face, narrative: Its aging players: The same real God . . . .” Everything is the same. “Urania’s Mirror” is about the concept of identity and about the Urania cards that


must be handled carefully, with white cotton gloves. Underlying it is the concept of being careful with the research and things of value—you can’t really touch things of value. “Urania’s Mirror” also reinforces the identities I talk about in Black Swan, in terms of being a southern black girl and what’s proper and what’s not. Everything white cotton gloves represent comes out in the poem.

Lucille Clifton says, “What did I see to be except myself,” or June Jordan says, “My name is my own, my own, my own” in the poems “won’t you celebrate with me” and “Poem about My Rights.” At a reading, Jasper Bernes—who studied poetry here at Cornell—read this line: “The universe, that cancer of the hallelujah.” It hit me as truth.

I see art being made in nature all around me, and this fascinates me. I’m interested in the connections between things, but I look at the connections by looking at the gaps. That’s what Open Interval is all about—looking at the gaps between things. Like in making lace, it’s the spaces that make the pattern. The lines that struck me in “Urania’s Mirror” were, “Take this compact–: Hold it up–: Squint until you recognize the light–:” What were you thinking? I was thinking, again, about beauty and Helen and the way she’s reflected in the mirror. The lesser notion is, if you’re the little southern girl or debutante, you have your compact: “Am I greasy/shiny?” That’s so ingrained in me from being raised in the South. The line is about holding the compact up and looking at oneself, recognizing all that goes into identity. What’s your favorite poem in Open Interval? Oh, that’s hard. I think it would be one of the “RR Lyrae” sonnets. “Will” might be my favorite. Why? Because it’s a love sonnet, and I love playing with the form. The “RR Lyrae” sonnets in this book have open intervals in them, and they have lines where I’m in conversation with other poets. I’m looking at the things that are behind the gaps: questions about identity like those raised when

I have a line in another sonnet that says, “One day I realized the space in everything is God.” Looking at the sky, the universe, and the space between the stars for a sense of interconnectedness is what I was conveying. I like that the poem ties everything back to the Earth. It has the line, “The universe, that cancer of the hallelujah,” in it that I borrowed from Jasper Bernes. What poetic techniques have you developed to help us, the readers, see what you’re trying to portray or see what you’re seeing? I borrow a lot from everyone. The first poem in the book is a bop poem, a form created by Afaa Michael Weaver at a Cave Canem workshop retreat. The way I’ve been writing the bop is a variation on his. I begin with a six-line stanza that establishes a problem or complaint. Then there’s an eight-line stanza and another six-line stanza. The eight-line stanza expands the complaint, problem, or issue. The last six-line stanza, if there is a resolution to be found, moves toward it. A refrain occurs between the stanzas. The refrains are usually taken from black music or black folk speech. I never want my poems to feel like I’m sitting in a room talking to myself. I try to impress

Joo Young Seo ’12

Urania ’s Mirror Here, before your face,

Whoever believes the mirrored world, shortchanges the world.


—Mona Van Duyn

its aging players: It took almost the same real God,

one hundred fifty years

the same false, the bull up there winking,

of charting heaven before the first celestial cartographer thought


to draw the night sky

his great red eye,


but you don’t see your wide black eyes

You handle these cards with


such care, reserved

Take this compact—:

for what is rare, precious: You are not allowed

Hold it up—: Squint until you recognize

to bring a pencil,

the light—:

to touch them; you must wear white cotton gloves.

Why Poetry? For as long as my memory goes back, I’ve written poems. For as long as I’ve had language, I’ve written poems. The first poem that I remember getting recognized for was in elementary school. I wrote a poem about a space shuttle launch. My teacher gave it to the principal, and they read it over the intercom when they did the morning announcements. My name, “Lyrae,” in Latin means “of lyric poetry.” I feel as though it was placed upon me. 52

this upon my students, as well. I’m engaging with the world and communities outside of myself. I’m in conversation with someone else. I love the bop in part because it includes lines from somebody else talking to me as I’m making the poem. In the bop, “The North Star,” the line I use, “Yes, the springtime needed you. Many a star was waiting for your eyes only,” is from Rilke. Have you created poetic techniques with your use of punctuation or word placement? People talk about the punctuation in Open Interval, because I use em dashes with colons. I was definitely in conversation. I always want to be in conversation with Emily Dickinson, for example: “I cannot live with you. It would be life.” In Poem 640, she says, “And that White Sustenance— / Despair.” I was writing Open Interval when I was hired at Cornell. I got here, and colons started showing up in my work in a way that they had not before. I think that’s Archie Ammons’ influence and my colleague Alice Fulton, with her double equal signs. Being in the building, I felt like I was writing in the ether. In Open Interval, I was also trying to be not only in conversation with literary people, but also across borders with mathematics and science. The dash and the colon together make interesting things. We learn about chemical bonds, like H2O, in chemistry classes. I wanted the dashes and colons to make the bonds between words more noticeable. Is the book, Open Interval, itself a technique? I was thinking across boundaries. Open interval is both a musical and mathematical term. The notion of bounded infinity is useful, because it’s everywhere. People have arguments about when life begins and ends. Life is an open interval. The universe is an open interval. God is an open interval. Faith is an open interval.

I was awarded an Appel Fellowship in 2011, which allows me time to work on The Coal Tar Colors, a new book. One of the things that’s driving the book is a painting by Eldzier Cortor called “Southern Gate.” It’s an African American woman nude in the backdrop of mountains, with a crumbling gate behind her. The crumbling gate behind her makes me think of academia. I don’t think that was intended, but that’s how I see things. A mockingbird is on her shoulder, talking to her. So I’m also thinking about birds.

Identity, in the book, is an open interval: “What did I see to be except myself?” “My name is my own, my own, my own.” I was thinking, someone’s identity is found in the open interval between body and name. How do you want your poetry to contribute to the world? I want people to read the poems—and then want to find and read the poems that my poems are in conversation with. I feel like something strange has happened now that people don’t read.

That wasn’t long ago. What has happened that people can’t name a working poet? I want people to read poems, because poems are beautiful and amazing, and they can point us toward what it means to be human and to see ourselves—to go back to that mirror.

` people/?id=64

If you ask people on the street to name a living poet who’s writing now, they’d have a harder time than people would have had in previous times. Edna St. Vincent Millay was like Angelina Jolie.


Single Molecules Trapped for Study Michelle D. Wang PHYSICS

To help us understand some of life’s mysteries, we have developed precision optical instruments and techniques to look at important molecules in biology—particularly DNA molecules and their associated proteins—one molecule at a time. A Focused Beam of Light When we highly focus a laser beam down to an extremely tight spot, it becomes very powerful. That tight spot becomes a trap for particles as small as a micron in diameter. If we move the trap, the particle follows. We can see the trapped particles, but at this point, we can’t directly visualize molecules because they are around 10 nanometers. So we attach a molecule to a particle, trap the particle, and use the trapped particle as a handle to sense what’s happening at the molecule and control it. More specifically, we have developed ways to exert controlled force and torque on the molecule and precisely measure its response.

Our work opens up a new dimension for studying deoxyribonucleic acid (DNA), motor proteins that travel along DNA, and other motor proteins. Our particular ability to exert and measure torque has received tremendous attention, and many labs are duplicating it. Looking Inside a Cell Nucleus Take a look. If we were to line up end to end all the DNA molecules that are inside the nucleus of a single human cell, the lineup would be more than one meter long. But a cell nucleus is only a few microns in diameter. A micron is one millionth of a meter. All those DNA molecules are squeezed into this very tiny space, all tangled and twisted up.

That’s a lot of compaction. And a profusion of topological activity accompanies the compacting. What’s more, biological molecules are all about motion. Some can translocate and rotate along DNA like tiny “molecular motors.” So we wanted to study the issues of topology, torque, and rotation, all at the same time. A DNA molecule is a right-handed double helix. Any motor proteins translocating along the DNA usually rotate around its backbone. As they move forward, we want to track their motion relative to the DNA. The ability to see all of this movement is immensely important to understanding what’s going on inside cells.

Research in Progress

Few techniques exist to do this consequential work. Biochemical techniques are typically indirect and deal with very large numbers of molecules at the same time. These techniques therefore sometimes miss important behaviors of DNA-based molecules.

T7 Helica se

Why Do We Care? The basic packing unit of DNA is a DNAprotein complex called a nucleosome.

Wang Lab

I have a large number of extremely talented students and postdocs working with me in my lab. They make a significant difference in my research. They make the research enjoyable. We observed the precise mechanics of how individual subunits of helicase—enzymes that travel along one side of doublestranded DNA—coordinate and physically cause the DNA unzipping mechanism. We were able to manipulate single DNA molecules to watch what happens when helicases encounter them and how different nucleotides that fuel the reactions affect the process.



We specialize in a technique called optical trapping, recording data from single molecules using a focused beam of light to “trap” microparticles attached to the molecules. This area of work is so big that we can work on many different problems. We can look at DNA compaction, transcription— genetic information being copied from DNA into ribonucleic acid (RNA)—and much more. Our work may enable solutions to problems in both biology and materials science.


location within one to two base pairs, or DNA units.

Conventional optical trapping is used to exert and measure a force on a biological molecule and monitor the molecule’s displacement. But we also wanted to be able to see rotation and torque. No one had demonstrated that optical trapping could be used to see the rotation of biological molecules and measure the torque exerted on and by them. To solve all of these problems and tackle the work, we developed angular optical trapping. It’s very direct and allows us to monitor rotation and torque. We literally watch how the molecules move and how they generate torque, and we can use torque to manipulate the molecule. Gene Information Storage and Retrieval In a very dense cell nucleus, tightly packed with a DNA-protein mixture, figuring out where a protein is bound on a DNA is essential to understanding how a gene is stored and how genetic information is retrieved. Our lab developed another technique, based on DNA unzipping, that allows us to see where a protein is bound on the DNA. We grab the two strands of the doublehelix DNA and pull them apart, like unzipping a zipper. We can unzip the DNA very easily until something gets caught in the “zipper.” We realize, “Oh, it’s likely a bound protein.” We have to unzip harder to pop the protein off, which is a great way to verify that it’s actually a bound protein and where it is. We can precisely determine its

Generally, DNA that is packed in a nucleosome cannot be accessed by motor proteins that might need to, for example, read the genetic information stored in the DNA. The interactions of DNA molecules with many different types of protein molecules are exceedingly significant to defining where a nucleosome is on a DNA template, which in turn determines whether a motor protein can access that part of the DNA. If the motor proteins cannot access a certain region of DNA, then the corresponding gene may not be expressed. The fate of that cell may be different from the fate of a cell where that gene can be expressed. So even with the same genetic makeup, the cells may actually develop differently, based on whether the gene is expressed or not. The Basic Search and Where It Leads We investigate questions relating to gene expression and regulation: how DNA is packed and how that packing alters the ability of enzymes that need access to the DNA for information decoding. Our work is basic research. It’s well understood in the scientific community that in order to find cures for a number of cancers, we must understand how cancers can get started. Understanding basic mechanisms lies at the heart of the development of a cure. How our research can be applied is unfathomable. Solutions to a multiplicity of problems may have a basis in this research. A variety of foundations have funded our work, including the Keck Foundation, the Beckman Foundation,

Wang Lab

Why this Research? I want to understand how nature works. Gene expression and regulation is very fundamental. I want to know how to understand it in a way that takes advantage of computing and precision instruments. We can exploit our . We can do modeling. We can create techniques and instrumentation to see what has never been seen before. And .


Research in Progress

Simultaneously measuring force, extension, torque, and angular orientation

Damon Runyon Cancer Foundation, the Sloan Foundation, the National Science Foundation, the Howard Hughes Medical Institute, and others. As a Howard Hughes Medical Investigator, I can explore whatever I want. I’m looking at different ways to study similar phenomena and more. Now we do single-molecule studies and work on theoretical modeling. Next, I want to look at the cellular level and see what’s going on in the complexity of cells.


We have developed an angular optical trapping instrument that permits simultaneous and direct measurements of force and torque for concurrent observation of the tensile and torsional properties of biological molecules over broad ranges of forces and torques. In an angular optical trap, four signals are simultaneously and directly measured: torque, rotation, force, and position, all with high spatial and temporal resolution. This wide bandwidth is also well suited for detection of highly kinetic processes. To ensure controlled orientation of the trapping particle and its specific attachment to the molecule of interest, we nanofabricated quartz cylinders that proved to be ideal trapping particles. Using these cylinders has dramatically enhanced the precision of torque measurements, making the angular optical trap a powerful tool for biological torque measurements. This instrument is opening up new possibilities for experiments on biological molecules, many of which are known to generate rotational motions and work against topological obstacles (e.g., topoisomerases, RNA polymerases).


Patricia Kuharic

Watching a Molecule Perform A Conversation with Scott C. Blanchard PHYSIOLOGY AND BIOPHYSICS, WEILL CORNELL MEDICAL COLLEGE

The most effective disease treatments rely on getting tiny molecules inside the body to find and stop individual molecules from functioning.

Finding better drugs for devastating diseases has taken many different forms. You concentrate on single-molecule imaging. How might your approach achieve superior therapeutics?


With single-molecule imaging, we try to understand the mechanisms by which known drugs operate. These studies reveal that many of the most effective drugs work by preventing or altering the way molecules move when they function. BLANCHARD:

We are developing fluorescence techniques for single-molecule imaging that will allow us to see more clearly and rapidly the

nature of molecular movements, their relationship to function, and how they can be targeted for therapeutic purposes. A more precise understanding of the mechanisms of known drugs will ultimately inform the development of newer, more effective ones. What does your strategy for seeing what a drug does entail? Let me first talk about the motivation. More than half of all known antibiotics target the bacterial ribosome, the engine that makes all the proteins in the cell. We

use them to treat eye or chest infections. They are so specific and effective at blocking the functions of the bacterial ribosome that we can take them orally, and they only block the bacterial ribosome and have no effect on the ribosomes in our own cells. This amazed me as a student, and I wondered why diseases such as cancer couldn’t be treated in a similar way. A key motivation for the work in my group is the notion that we might be able to do so, if we had a better understanding of how enzymes such as the ribosome work.

Research in Progress

My graduate work at Stanford in structural biology and biophysics focused on understanding three-dimensional structures of molecules responsible for various life processes. Our current era of structural biology has been successful because we can obtain incredible pictures of biological molecules in frozen states that give us snapshot views into the molecular basis of life. These pictures are very important for understanding mechanism, drug activity, regulation, and more. But the pictures are static, and we understand more and more that life processes are driven by constant motion. My interest in single-molecule imaging grew out of the need to figure out ways to obtain dynamic views of functioning molecules: how they normally move when they work and how regulation affects their movements. I wanted to know how the specific class of antibiotics that I worked on at the time affected protein synthesis catalyzed by the ribosome. The question was, how does a drug so small affect a giant molecular machine, the ribosome? It’s a real David versus Goliath story: the drug is about 5,000 times smaller than the ribosome. I wanted to understand how this worked at



When I first became involved with single-molecule imaging as a graduate student in 1998, I was immediately struck by how potentially powerful it would be to watch molecules in real time, like watching a movie. To bring this science to a medical institution is tremendously important. We need to move the technique from a niche biophysical approach that can only be performed by specialists to one that can be applied to a broad range of medically relevant systems by students and medical scientists in order to unleash its true potential.


a molecular scale, but the tools of the day weren’t adequate.

Where are you currently in technique development?

But you needed tools to see. To see the ribosome in motion, we needed a time-resolved method that gave us both structural and kinetic information, and most importantly, we needed a method that could somehow allow us to watch individual molecules at work, because otherwise we would just see the average behaviors of the system and lose all relevant information about motion.

The single-molecule fluorescence imaging technique is built on a foundation of research that’s been in development over a very long period of time. Our technique gets better and better every year as the individual components of the complex instrument we use are advanced. This very sensitive technique demands very fast cameras and exceptionally stable

The question is whether there is something to learn from the story of how antibiotics are used to treat bacterial infections. Capitalizing on a wave of similar pursuits going on at Stanford at the time, we decided to pursue the development of single-molecule fluorescence imaging techniques to eavesdrop on individual ribosomes in action. Fluorescence had long been known to be one of the most sensitive and reliable ways of looking at molecular motions. Foundational studies on simpler systems had already shown the proof of principle, so off we went. And you have continued to advance the research on how to see the action?

lasers. We don’t develop cameras or lasers. We rely on industry for these, so we’re constantly looking for the right components to assemble into our own homemade microscope for the best performance, like building a hotrod. It requires all the parts to fit together just right: fast cameras, stable lasers, good temperature control, long-lasting fluorophores with good signals, extended time-performance microfluidic devices, and computers for rapid, nonbiased data analysis.

The real challenge was figuring out how to place the fluorophores—and the right fluorophores—in places that allow us to see the moving parts of the ribosome. How do we place fluorophores exactly where we want them, like a hidden camera on the wall, without affecting the ribosome’s normal functions? How do we make the fluorophores bright enough and last long enough to see what we want to see?

Our key contributions in the field have been technique development for site-specific labeling of biological molecules, the development of new fluorophores that can be used to look at their motions with greater and greater resolution, and the development of robust computational tools to treat the data obtained. Understanding how to interpret the data has been an active area of research for us, as well as a challenging area of research for the field, and we have very talented students working on this.

It took about five years to overcome these challenges—to get our proverbial foot in the door and see the ribosome at the singlemolecule scale. We have many more tools at our disposal now, and we are just beginning to be able to see the true nature of this molecular machine’s intricate dance for the first time.

We are also slowly developing nextgeneration microfluidic devices to make single-molecule imaging faster, more generalizable, and more available to the masses. With the help of the Cornell nanofabrication facility in Ithaca, we build our microfluidic devices in-house. Few researchers can do this.

Blanchard lab

Patricia Kuharic

Why this Research? I tried banking—my first degree is in economics. I tried engineering. I stumbled upon science, and I love it. I was an MD/PhD student when I first started out in science, and discovered that I like basic research more. I wake up thinking about biological systems. I go to bed thinking about them. We rely so heavily on drug treatment for cancer, lymphoma, and various infectious diseases. To make advancements toward the treatment of diseases like these is worth dedicating my career. All the antibiotics that we rely on—found in the dirt or tree bark or coral in the ocean—work, but we don’t know why or how they work from a physical perspective. 64

It’s a significant shortcoming that more people are not focused on this area of research—a very important frontier— where we have the potential to control disease in elegant ways. Of all the areas I could apply my life to, this is it. When my mother was taking chemotherapy for cancer, I thought how sad that we don’t have a more elegant solution to cancer therapy, like an antibiotic. You get sick, you take a pill, you’re off it three days later, and you feel beautiful.

Research in Progress



Blanchard Lab

Our recent paper in Nature Methods (2012.9) describes a potentially generalizable method for greatly improving the performance of commercially available fluorophores. In Science (2011), we describe the first crystal structure of the bacterial ribosome in hybrid states of tRNA binding. Our single-molecule fluorescence methods revealed that this configuration of the ribosome is important to function and showed us how to stabilize it for crystallographic studies.

All to be able to see single molecules in motion? Single molecules in motion—and to understand how antibiotics work and regulation takes place. You received an NSF career grant a couple of years ago. How does that research relate to being able to see single molecules perform? The goal of the field is to develop these techniques to be robust enough that researchers can simply watch biological processes inside cells, which will require new instrumentation. The typical biochemistry experiment takes place in a test tube with many microliters containing billions or even trillions of molecules simultaneously. What scientists get out of this is information about the average behavior of the molecules they interrogate. Using our system, in contrast, we can get quantitative, real-time feedback—all the information we need from 1 to 1,000 molecules. This is on par with the actual number of molecules controlling life inside the cell. The goal of our NSF grant is to move the field away from traditional model systems toward more biologically and medically relevant systems. Almost everything we know about the mechanism of translation has been built upon studies of the highly simplified machinery of the bacterial cell.

It’s been an incredibly powerful and informative system, but we now need to extend ourselves to investigations of more complex systems up to and including the human ribosome. These systems have been impenetrable because it’s difficult or impossible to obtain enough material to investigate them (think trillions of molecules in a test tube). Now that we have the platform for investigating these systems at the single-molecule scale, the stage is set to move toward the types of systems that ultimately need to be interrogated for the treatment of human disease. Since we’re talking about applying techniques of single-molecule imaging to medical problems, how would your research apply to cancer? We would all like to be able to treat cancer like we treat a bacterial infection—swallow a few pills for a week and be cured. What we do today is far more taxing and painful. Essentially, we try to kill everything that is growing rapidly and hope that the body and the immune system figure out the rest.

the bacterial ribosome can perform so effectively when taken orally. It turns out that the ribosomes inside cancer cells are also physically and functionally distinct from ribosomes inside most normal cells. Is there the potential for a smallmolecule therapy for treating cancer? Only time will tell, but we hope that the techniques we are currently developing will help us figure this out. Few, if any, biophysical measurements of human translation activity have ever been taken. Wouldn’t it be great if we could advance the field of single-molecule imaging to a point where we could make biophysical measurements of ribosome activities within healthy and diseased cells, and then use this as our platform for developing or discovering drugs that kill the bad cells and not the good ones?

` blanchard/lab

The question is whether there is something to learn from the story of how antibiotics are used to treat bacterial infections. The bacterial ribosome and the human ribosome are physically and functionally distinct. That’s why antibiotics targeting


Patricia Kuharic


The Optical Biopsy Douglas S. Scherr


We want to create a better way to look inside the walls of a bladder for cancerous cells, eliminating the constant need for biopsies and saving the patient so much morbidity.

Woes of Bladder and Prostate Cancer A person is diagnosed with bladder cancer. It may be superficial bladder cancer, but it won’t matter. Copious pain, bleeding, risk of infection, risk of perforating the bladder, time off from work, and expense will be the patient’s ordeal. Every three months of the first year after diagnosis, the patient must get a cystoscopy, looking inside the bladder with a camera. When we look inside the bladder, oftentimes we see little red spots and areas that ordinarily would be of no concern. But in a patient with bladder cancer, we must be concerned with everything,

because the naked eye cannot always tell whether the tumor is benign or malignant. So, the patient must get a biopsy every three months, particularly early in the disease course. This is a lifelong experience— lifelong surveillance. Bladder cancer—more than any other cancer— is the most expensive disease to care for over the course of a lifetime. Can we find a better way to see not only inside the bladder but also beyond the lining of the bladder, inside the wall at a microscopic level, without taking a biopsy, saving the patient from so much morbidity?

A Burning Aspiration I’m a urologist, and my specialty is urologic oncology. I care for patients predominantly with bladder, prostate, kidney, and testicular cancer, where I’m removing mostly bladders and prostates, and some kidneys and testes. I see a large volume of patients with these diseases. The treatment is multifaceted, including surgical therapy, chemotherapy, and radiation therapy. Our multidisciplinary care team includes the medical oncologist, radiation oncologist, pathologist, radiologist, and urologist. At the heart of the treatment is a lot of basic science and translational research.


Research in Progress

Douglas Scherr

of H&E Comparison oton Images h ip lt u M d an These images are from the bladder. The colorful ones are the multiphoton microscopy (MPM) images, and the purple ones next to them are the H&E images. An advantage of multiphoton imaging over H&E imaging is that we can do what is called Z-stacking. When we take a biopsy, we are taking a frozen minute in time, looking at that section. Multiphoton imaging lets us start with the superficial and image deep without moving the scope by adjusting the laser. This is in two dimensions, but effectively we are seeing it in three dimensions where we can Z-stack the images. We have a stack of about 50 images, and we can focus in and out and actually see in real time where tumor cells are and if they are superficial or going deeper into a layer.



Multiphoton microscopy allows imaging directly through living tissue. Multiphoton endoscopy is like a multiphoton camera that can be inserted into the body to do real-time imaging, averting the need for so many biopsies. The multiphoton endoscopy project has an interdisciplinary team of clinicians, physicists, biomedical scientists, and engineers. Watt Webb and Chris Xu, Applied and Engineering Physics, are at the center of the project. With a team in the College of Veterinary Medicine, including Alexander Nikitin in Biomedical Sciences, Warren Zipfel in Biomedical Engineering, and a group of about 15 clinicians at Weill Cornell, we have brought together a fascinating group from Weill Cornell and the Ithaca campus.


But until we are able to cure all of these diseases, particularly prostate and bladder cancers, at a 100 percent cure rate, we need to focus also on understanding the biology, natural history, and genetics of these diseases. This has been my burning desire—to understand these facets of the diseases with more clarity and to discover how we can apply what we learn in the lab to clinical medicine and patient care. Seeing to Cure About six years ago, I was at a meeting in Ithaca, where I met several biomedical engineers and applied physicists. As we began talking, I explained these same problems. How do we see microscopic findings without a biopsy? The optical biopsy has always been the Holy Grail of futuristic medicine. Can we look inside someone’s body without taking a biopsy? This is where my interest in multiphoton endoscopy (MPE) began, and we struck up a collaborative search. How do we create something to put inside a patient and perform in vivo multiphoton

imaging of living human tissue. Multiphoton microscopy (MPM) is the brainchild of Cornell professor Watt Webb, Applied and Engineering Physics. It has been utilized mostly for ex vivo imaging, like any microscopy—you take a piece of tissue out and look at it. How do we create an endoscope? The Collaborative Search Researchers at Ithaca and Weill Cornell are working in parallel on the problem. The engineering group in Ithaca is fabricating the actual endoscope: the optics, lenses, and lasers. At Weill Cornell, we are developing the paradigm in humans, creating an atlas of multiphoton imaging to define its extent and understand its imaging properties. We are creating a whole new field of microscopy that allows physicians to see the same features they would see using H&E staining, but with multiphoton imaging to see if the two methods are in sync with one another. Physicians and pathologists are used to looking at H&E staining, the standard for biopsies. No one has this experience looking at multiphoton imaging.

Jason Koski/CU

The Watt Webb and Chris Xu, Applied and Engineering Physics Team

Jason Koski/CU

Multiphoton Endoscope Prototype

WCMC Urology

Why Urologic Oncology? Patricia Kuharic

It’s a tremendously exciting time to be a doctor. Only a few fields could allow me to put together diverse collaborations, working with so many different disciplines all trying to achieve the same goal of improving patient care. Translational research is the ultimate gratification— helping patients, improving technology, and improving science. At the same time, it allows me to be a lifelong student. Every day I feel like I’m still a student. The minute I’m no longer learning, it’s time to hang it up.


Research in Progress

What are we seeing when viewing multiphoton images? In our blind testing, a group of pathologists examine several different organ systems—the bladder, colon, breast, thyroid, ovary, and lung. We take an excised piece of human tissue, image it using both the H&E and multiphoton methods. We then compare the two to see if we can be diagnostically as accurate using multiphoton diagnostics as we can with H&E.

One of our major goals in multiphoton imaging is to specify stage. Can we determine depth? We discovered that this imaging can penetrate about 500 microns, which is enough to get through the mucosa and into the lamina propria layers, but not into the muscle. Diagnostically, this is invaluable.

This is not the first attempt at doing an optical biopsy. What is new is that MPE is the only technique that gives a true H&E-quality histologic image without taking a biopsy. It is the closest technique to the gold standard. at totally different cell types. Sushmita Mukherjee, director of the multiphoton microscopy core facility at Weill, has been instrumental in fine-tuning the optical parameters of the multiphoton scope and has helped in the development of the multiphoton atlas. We developed a technique called a punch biopsy of the tumor. We take a tiny needle with a hollow core and push it into the tissue to remove a tiny cylinder of tissue. This way we are only looking at that one cylinder spot. We shave a slice of the cylinder and image it. This technique ensures that we are imaging the exact spot under both H&E and multiphoton methods. Seeing Cancer Stage and Grade With cancer, we want to figure out stage and grade. Stage refers to how deep into the bladder wall the tumor invades. The bladder is made up of several layers. The first layer, mucosa, is where cancers start. The mucosa is lined with transitional cells. A cancer of the lining of the bladder is called transitional cell carcinoma. The second layer, lamina propria, consists of many small blood vessels and lymphatic

ell Weill CorCnore facility

Determining stage means finding exactly where in the wall the tumor invades. About 70 percent of bladder tumors are superficial, meaning they are confined to the lining of the bladder, but the other 30 percent invade deeper into the bladder wall.

We would be able to see if the tumor has penetrated the first layer of the bladder, and this in part will give us the stage. To find out if the cancer has gone into deeper layers, we scrape away the top part of it and reimage. The biologic aggressiveness of the cells is the cancer’s grade. Bladder cancer can be low grade, meaning the cells look only slightly abnormal, and they do not have the propensity or the ability escape the bladder. Alternatively, they can be high grade, where the cells look bizarre—very unusual in appearance. These cells do have the ability to escape the bladder and metastasize. Multiphoton Endoscopy at the Proving Ground The standard method for diagnosing tumors and cancers is the biopsy— taking a piece of tissue and processing it. We want to do this whole process using a multiphoton microscope—the optical biopsy.


Jason Koski/CU

Overcoming a Major Challenge It’s a major challenge to compare an H&E image to a multiphoton image. First, we need to be sure that we are looking at the same exact spot of tissue. Looking under a microscope that has 40x magnification, if we are one micron off, we will be looking

vessels. Underneath is a muscle layer and on the outside is fat.

Sushmita Mukherjee (l.), Biochemistry, heads Weill Cornell’s multiphoton microscopy core facility. The lab has two multiphoton microscopes. We can take a specimen from a patient, put it on a multiphoton microscope block, and image it with no processing.

diagnoses. The work at Weill Cornell and Ithaca will converge, and we can begin to look at endoscopic images. We’re very close. We have published several papers, and the accuracy of our pathologists in diagnosing tumors using MPE is quite impressive. We are completing an animal prototype to be tested at the Cornell veterinary college. Chris Xu, Applied and Engineering Physics, and Alex Nikitin, College of Veterinary Medicine, are working closely to do the initial small animal imaging and ultimately larger animal imaging.

` dsscherr

We are developing multiphoton endoscopy as a proof-of-principle concept. Once the endoscope is ready, we will have justified the MPE as a legitimate method for assessing histologic images of tissue and making



Optical Tools for Seeing How Disease Changes the Brain Chris B. Schaffer BIOMEDICAL ENGINEERING

How do injuries to small blood vessels in the brain contribute to brain disease? Why don’t axons regrow after a spinal cord injury? What is the relationship between impaired brain blood flow and Alzheimer’s disease? To find out, we develop tools and techniques that let us see what is happening to individual cells inside the brain during disease development in animal models. Injuries to the Head We have all heard of athletes, particularly hockey and football players, who received numerous head injuries during their careers. These injuries are believed to lead to degenerative brain diseases similar to Alzheimer’s disease. Some of these athletes died at an early age and postmortem diagnoses pointed to vascular injuries as a potential cause of the degeneration. We know that such injuries can cause a range of disorders in the brain, but we do not know how.

The Right Tools Whether it’s Alzheimer’s disease, epilepsy, small strokes, brain cancer, or spinal cord injury, dysfunction results from a disturbance at the cellular level inside the central nervous system. To see at this scale inside the brain—actually observe how individual cells are affected, as well as where and when after an injury or in disease development—is very difficult. Yet it is essential to see the cellular behavior in order to comprehend the mechanisms that lead to dysfunction. A major focus in my

lab is to understand the cellular-level changes that occur in the central nervous system and lead to these diseases. Because of the technical challenge of doing such experiments, my lab develops tools and techniques and uses them to answer key scientific questions. We focus about one third of our effort on development of novel techniques and two thirds on answering questions about neurological disease states, often using advanced tools that we develop.


Research in Progress

Disturbances in the Brain Medical scientists have known for some time that injury to small blood vessels in the brain is associated with cognitive decline and dementia in aging humans. They have found in postmortem examinations that people who exhibited more severe cognitive decline had more small-vessel injuries. This kind of clinical data, however, does not help us to understand the mechanisms by which small-vessel injuries lead to brain cell dysfunction. Getting at the Root of Microvascular Stroke To uncover what’s happening at the cellular level after a microvessel injury, we need an intact circulatory system carrying blood through the brain, so we do our experiments in animal models—mostly mice. We cause an injury by clotting or hemorrhaging a



We image the brain of live anesthetized mice after removing a section of the skull to gain optical access to the brain. The mice recover from the surgery, and we can come back and reimage the same animal over time after an injury. We can find the same blood vessel that we hemorrhaged. We can find the same dendrite that sits next to that blood vessel, and we can even find the same dendritic spine—the micron-sized structure where neurons communicate with each other. We can find a one-micrometer region located within the brain of the mouse and tell whether it changes day to day. We’ve found that after some kinds of brain injury, new dendritic spines form and existing ones die at a faster rate than in controls. This random rewiring could, even in the absence of cell death, lead to cognitive dysfunction.


little vessel and then directly observe how the injury affects the cells in the brain in order to help us understand how brain cells become dysfunctional after such small vascular injuries.

however, the mechanism that leads to dysfunction appears to be not so simple. We were surprised to find that when we put a small hemorrhage in the brain, the nearby cells did not die. The neurons around the

In addition to studying the effect of small vascular lesions, their link to Alzheimer’s disease, and cellular dynamics after spinal cord injury, my lab also has projects that focus on studies of epilepsy, brain cancer metastases, blood cancer, neural prostheses, and the development of next-generation tools for biomedical research. We produce vascular injuries using intense pulses of light that are fantastically short in duration, about 100 millionths of a billionth of a second long. This short pulse of light can be used to injure the wall of a targeted blood vessel. For larger injuries, this causes the vessel to rupture and produce a small hemorrhage. With a more subtle injury, the bleeding is limited, but the injury initiates clotting and a blockage is formed in the vessel. We then use laser-based optical imaging techniques to see how individual cells are affected, how they die, how they lose function, what new cells invade the injury site, and how they interact after the microvessel injury. We can identify individual cells and subcellular features and track how they change over time, from minutes to months after the injury over a spatial scale of about a millimeter and with micrometer spatial resolution. Our goals are to reveal the mechanisms underlying these brain diseases and identify therapeutic targets to treat the problem. A Closer Look at Small Hemorrhages Recently my lab began looking keenly at small hemorrhages. We and others have previously found that occlusion of small blood vessels causes the death of nearby neurons and other brain cells. For small hemorrhages,

small hemorrhage were unharmed and remained so for weeks, as long as we watched. Inflammatory cells invaded the area near the hemorrhage, but they did not seem to do anything catastrophic. So whatever causes the dysfunction does not kill the cells. In preliminary data, we do see that there is an elevation in the rate at which the neurons near the hemorrhage change their pattern of wiring to each other. This random or semi-random rewiring of the neurons in the brain may disrupt the brain’s normal function and is a potential cause of cognitive dysfunction following small hemorrhages. Only by Seeing We could have made this discovery only through an imaging technique—we need to see an individual synapse before producing a little injury. After we produce the injury, we need to see it again. We have to see it again two days later, two weeks later, and see if the same synapse is still there or if it went away or if a new one formed, and at what rate synapses are lost and gained. We need the combination of advanced optical technologies for imaging and the ability to manipulate with cellular resolution in a live animal, having clear hypotheses about what structural and functional changes may underlie disease, in order to make advances.

Research in Progress

In very recent work, we have found that small vascular occlusions occur spontaneously at a much higher rate in mice with Alzheimer’s disease than in normal mice. Together with our results suggesting that small strokes can drive amyloid-beta accumulation, this suggests that microvascular dysfunction and the pathology of Alzheimer’s disease form a vicious cycle that mutually exacerbate each other.

Linking Small Strokes and Alzheimer’s Disease Researchers have shown clinically that Alzheimer’s disease and microvascular injuries in the brain are often associated with each other—they are comorbidities. Is it because microvascular injuries trigger Alzheimer’s or does Alzheimer’s pathology cause microvascular injuries—or both, or neither? Alzheimer’s disease is a clinical diagnosis. It could be that, in order to exhibit enough dementia to be classed as having Alzheimer’s disease, two independent disease pathways are affecting the brain.


Why this Research? Short-pulse laser systems have been a consistent part of my research career. As an undergraduate, I worked on designing and building laser systems that produced femtosecond-duration laser pulses. These pulses were as short as 10-14 second in duration! I studied the physics of the interactions between these pulses and materials as a graduate student. When I became a postdoc, I realized that all I had studied and done could be applied to create what is essentially a fancy laser scalpel for biomedical research, and perhaps it could be used as a surgical tool. I transitioned to biology-based research in order to apply some of the tools I had discovered. One aspect of my current research is to develop high-fidelity research tools for studying diseases in animal models, so that we can uncover the underlying mechanisms of disease. If we understand the mechanisms that lead to a disease, then we can identify therapeutic targets for drug development, surgical strategies, or a broad variety of medical approaches for treatment or prevention. 78

Research in Progress

Schaffer Lab

We label structures and cells of interest by injecting a fluorescent dye into the bloodstream, and all the blood vessels carry it, like angiography. We can add fluorescent dyes that give structural information about where neurons are, determine if cells are dead or alive, and even discern if the ability of a neuron to do its job has been altered as a result of disease. In the image, we visualize the amyloid-beta aggregates that are characteristic of Alzheimer’s disease (green) and the blood vessels (orange) in the brain of a live mouse.

To help sort this out, our experiments use mice genetically engineered to get Alzheimer’s disease to see if more Alzheimer’s pathology is triggered when we induce a small stroke. Alzheimer’s is caused by a buildup of amyloid-beta, a small peptide that is produced by neurons and cleared through the vasculature. It is the aggregated form of this peptide that causes neurons to lose function and die. We found that when we blocked a small blood vessel in the brain, this triggered the aggregation of new amyloid beta in the immediate vicinity of the clot. This makes sense, because if we clot a blood vessel, it’s like plugging the drain that amyloid-beta uses to get out of the brain. This leads to an increase in concentration of amyloidbeta, and therefore more aggregation. Our data suggest that cerebral vascular disease could be an initiating or exacerbating factor in the development of Alzheimer’s. Our finding is exciting because it suggests new preventative or treatment strategies for Alzheimer’s. We could target the vascular component of Alzheimer’s disease

independently of other strategies being developed to target the amyloid-beta aggregation or neural dysfunction. Building, Experimenting, and Getting Illuminating Results Creating the right tools to see is pivotal. For example, many studies of spinal cord injury in animal models were limited by the tools that were available, primarily behavioral assays and postmortem histology. As a result, incorrect or ambiguous conclusions were drawn. Researchers did not have the right tools to answer vital questions like, is an individual axon dying back or growing after an injury? Recently, we developed a technique to image such dynamics and quantify them. It turns out that, for the study of spinal cord injury, the technique development that was needed involved a surgical preparation, not a fancy optical tool. We needed a good way to do surgery on a mouse so that we could have optical access to the spinal cord and keep that access for days, weeks, and months.

With this new approach, we can image the dieback and potential regeneration of individual severed axons in the spinal cord of the same mouse over time and with micrometer-scale spatial resolution. Our technique will help us and other researchers in understanding spinal cord injury and developing therapeutic strategies. Seeing the Invisible Axons, the neurons that conduct impulses away from a cell body to other cells, in the spinal cord are wrapped with many layers of cell membrane called myelin. Researchers believe that the loss of myelin surrounding individual axons may be a factor in degeneration after spinal cord trauma and that this loss impedes regeneration. Myelin provides essential electrical insulation for the axon. Without it, the signal that is conducted down the axon tends to fail. In some spinal cord injuries, evidence suggests that the axons are not severed. They have lost myelin, and that’s why the conduction is not good. With poor conduction, the axons do not communicate well with the central nervous system about what the


Research in Progress

Schaffer Lab

Two-Photon Excited Fluorescence Microscopy We use an imaging technique called two-photon excited fluorescence microscopy, which was developed here at Cornell in the early 1990s in the lab of Watt Webb, Applied and Engineering Physics. It is a way to image fluorescently labeled objects in three dimensions inside tissue, while at the same time alleviating the problem of loss of contrast and resolution that is caused by scattering of the fluorescent light. The image shows a 3-D rendering of blood vessels in the brain of a live mouse.

muscles should do. In order to understand what happens to myelin and develop strategies to encourage remyelination, it is necessary to be able to see the myelin in a live animal. We have many fluorescent labels that can be used to mark the cells and axons in the spinal cord, but labeling myelin for imaging is very difficult, because the fluorescent dyes that can be added to myelin dramatically disrupt the structure. In recent work, we found that a nonlinear optical technique called third-harmonic generation (light of one wavelength is converted into light one third of the wavelength by interacting with the sample) is very efficient whenever we have a bold optical interface, like that


produced by myelin (mostly lipid) that wraps around an axon (mostly water-filled). Using third-harmonic generation, we can see inside a mouse’s spinal cord after a spinal cord injury and investigate how changes in the amount or structure of myelin occurs. We want to use this tool ultimately to study strategies for triggering remyelination in injured axons, moving toward therapies for spinal cord injury. The Rewards of Seeing Clearly Our research approach of developing tools that enable novel classes of experiments is especially rewarding when we can come into a field with a key unanswered scientific question. Often the question is obvious,

but the right tool is not available to study it. We can build a tool for the experiment, do the experiment, and get the whole field moving again.




Flies and Humans: Looking into Genetic Variation in Populations Andrew G. Clark


How do differences in DNA sequences in a population produce variability in complex traits? We hope to learn basic principles of this correspondence in flies and apply those to understand variation in disease risk in people. From Flies to Humans As an undergraduate I had the great fortune to be working in the lab where Margaret Kidwell was doing her early work that led to the discovery of the P element, the first transposable element used to manipulate the fly genome. I started out in the laboratory working with fruit flies, Drosophila melanogaster, because it was easy to set up populations that tested population genetic models. Drosophila also happens to be a terrific model organism for understanding many basic mechanisms in genetics and cell biology.

Years later, when I established my own research lab to study the population genetics of Drosophila, my group became interested in the problem of modeling fat storage in flies. We wanted to be able to predict— given information about the genetic composition of flies with diverse genetic backgrounds—which lines of flies would store more lipid. This work attracted the attention of researchers in human cardiovascular disease, since they were trying to model the genetic basis for differences in cholesterol deposition in blood vessels. So I got involved in this collaboration starting in 1996, and it has continued to this day.

Working with medical collaborators about 10 years ago, we sequenced a specific set of genes in the reverse cholesterol transport pathway of several people. These analyses really opened researchers’ eyes to the abundance and complexity of genetic variation in human genes. Among a sample of 71 normal people, there were 88 DNA sequence differences in just one gene! This work attracted the attention of Craig Venter and his colleagues at Celera Genomics, just as they were setting up to sequence the Drosophila and then the human genome. I did weekly consulting work for Celera for four years, until after they completed the human, mouse, and Anopheles mosquito genomes.


Research in Progress

This was quite an introduction to genomics! The very notion that it was possible to obtain a complete genome sequence— where every gene and every regulatory element was there, on the computer—was exhilarating. It provided a whole new way of thinking about many problems in biology, and we are still to this day just beginning to capitalize on the power of genomics approaches.

aggregate in families—that is, they occur in relatives at a higher rate than in the general population. This suggests that genes may be involved, and several approaches have been developed to try to identify the disease-causing variants in these genes. The ways that such variants are identified in humans and flies are very different, but the underlying principles of population genetics remain the same.

What Does Cardiovascular Disease Have to Do with Fruit Flies? My initial work with Drosophila melanogaster asked questions about variation in several phenotypes—observable characteristics of organisms—in the population of flies. We performed experiments to measure how genetic variation caused the observed phenotypic variation. Much of human genetics has the same primary challenge: we want to know the genes and their variants that make some people get sick.

It turns out that one of the most direct ways to map disease variants, called a genome-wide association (GWA) study, can be distorted by the fact that different people in the study may come from different ancestral backgrounds. This gave us considerable motivation to determine how well we could infer someone’s ancestral history based only on their DNA. These methods have also progressed enormously in the last 10 years—but before we get to that, we need to return to fly stories.

In the case of cardiovascular disease, we know that many types of the disease

Comparing Whole Genomes of 12 Fly Species While the human disease work was accelerating, the Drosophila community was also moving forward. It became inexpensive enough to sequence whole genomes that a group of us proposed to sequence the complete genomes of 12 different Drosophila species. These were species whose relationships were known—all 12 species had a common ancestor that lived about 60 million years ago. The idea was that we could compare two or more species and see where all the changes were as one species evolved to another.



We recently did a paper on a genetic system within flies called piwi RNA—a class of RNA like an immune system against transposable elements of viruses. The way it works is very intricate, and we see this! It is unbelievably beautiful in the way it allows the genome to collect bits of viral sequence so that the genome can remember and later fight that particular virus.

This is an immune system with memory. Nobody had a clue how it worked until only a couple of years ago. From an evolutionary perspective, it makes absolute sense. We understand many attributes of the system. It is deeply evolutionary, and so by studying this mechanism in flies, we learn about it in humans. And it hangs together beautifully.


An obvious and still very exciting question is whether we can infer which DNA sequence changes caused which changes in the organisms themselves—their appearance, behavior, and so forth. But starting with the simpler problem of just getting the sequences, aligning them, and seeing how changes fell on the tree of relationships among the species was a very exciting start. We learned, for example, that genes that encode proteins in seminal fluid are the fastest evolving genes in the flies, and that immune system genes are the next fastest. Evolution was driving these genes to evolve rapidly, and we have the tools now to understand why. Heading up the team

that pulled together this initial analysis, with students Amanda Larracuente and Tim Sackton and postdoc Nadia Singh, was a close match to the excitement of working with Celera Genomics. Obesity in Flies?! My team’s work on lipid storage in flies continued, and we even created lines of flies that were highly resistant to starvation. They achieved this trick by basically shutting down the mechanism that tells flies to stop eating when full. These flies gained three or four times as much fat, and they were even insulin-resistant, just like diabetic people. We were funded to pursue this problem of obesity in Drosophila, and in this one project, we sequenced the genomes of 92 lines of flies. We also measured expressions of every gene, enzyme kinetics, lipid composition, and many other metabolic attributes of the flies. The big challenge remaining is to model the regulation of metabolism across these 92 fly lines. The ultimate test is to be able to predict from the DNA sequence, when presented with a particular diet, how much fat the flies will store. We are already getting pretty good at this, but there is room for improvement, and at each step we learn how the models work. This ability to predict disease risk is a highly desirable goal in human genetics, as well, although because we cannot control the environment, the expectations for prediction in humans is much more modest. At least we hope to identify individuals with the greatest genetic risk, so that they might consider ways to change their diets or other aspects of the environment to minimize future disease risk. Stumbling into Tuberculosis Our work on Drosophila innate immunity attracted the interest of James Musser, a medical doctor who has an interest in genetic differences among humans in resistance to tuberculosis. Now, tuberculosis is an infectious disease, so you may think that whether or not you get TB depends only on exposure—but nearly everyone is exposed to Mycobacterium tuberculosis, yet only a few get TB. Tuberculosis is a good example of an infectious disease that

Research in Progress

nevertheless has a strong genetic component for susceptibility. In the end, those genes make a big difference in who stays healthy and who gets TB. The study we did was to sequence genes called toll-like receptors in a large collection of TB cases and in people who were healthy controls, but who were known to have been exposed, because they shared IV needles with the TB patients. We sequenced the toll-like receptors because these are genes—first discovered in flies—that play a key role in the initial recognition of infection. The parallel between the human and fly innate immune system is striking, right down to the fact that variation in these toll-like receptors results in differences in the ability to recognize that the host is infected. In fact, our study in humans found that toll-like receptor differences were associated with rather large differences in TB susceptibility. Genomic Imprinting across the Genome At the same time that the human and fly work was progressing, a talented graduate student in my group, Xu Wang, wanted to pursue genomic imprinting. I had written several theoretical papers on this subject, because genomic imprinting presents such a challenging evolutionary puzzle.

Genomic imprinting occurs when either the mother’s or father’s allele for a gene is shut down and not expressed. This is a puzzle, because if the mother’s copy of the gene is shut down, and the father’s copy is defective, then only the defective copy is expressed. We started with mouse crosses, relying on Paul Soloway, Nutritional Sciences, for help with the mouse work. Instead of sequencing DNA, we sequenced the complete collection of RNA copies made from the expressed genes (a method called RNA-seq). We were able to find variants between the mother’s and father’s copies of most genes, and so we could count up expressions of the maternal and paternal copies. Some genes only express one or the other copy, and these are the imprinted genes. We discovered a few new imprinted genes, and were bitten by the bug to try this approach on diverse organisms. Horses, Donkeys, Chickens, Bees, and Wasps Early on we made contact with Doug Antczak, former director of the Baker Institute for Animal Health at Cornell, who has researched horse-donkey hybrids for 30 years. He had been studying the placenta in horses and donkeys, including the situation when the fetus is an F1 hybrid (a mule if the mother is a horse, or a hinny if the mother is a donkey). Antczak was able to obtain samples of extra-embryonic tissue before it was implanted in the uterus. In horses, it is possible to sample the tissue that will become the placenta without any maternal contamination. (This is a very challenging dissection in mice.) The horse-

donkey system has proven to be very powerful, and we are finding many novel imprinted genes. It is a strong collaboration with many new discoveries awaiting us. We have since applied the reciprocal cross trick in several other collaborations in diverse organisms, including chicken, honeybee, a wasp, and the South American opossum. The simplest take-home message so far is that genomic imprinting is highly fluid, with the set of imprinted genes highly variable across species. At least in mammals, however, there is a small core set of 15 to 20 genes that is stably imprinted everywhere. By completing our comparative study of imprinted genes, we will be in a much better position to tease out why the genes that are imprinted have this odd manner of expression. The Importance of Ancestry in Assessing Genetic Disease Risk Right after the human genome sequence was published in 2001, the human genetics community was very eager to identify DNA sequence variants across the entire human genome. This started with a large collaborative project to identify and map haplotype variants called the International HapMap Project. Haplotypes are collections of nearby DNA sequence variants that recur in many individuals. The HapMap Project was a big success in identifying more than 10 million variants, and it also spurred commercial interest in developing cheap ways to determine the SNP genotypes of many positions in the genome for each individual.

? s ic t e n e G n io t a l Why Popu As an undergraduate, I did a double major in biology and applied mathematics. I always appreciated problems in science with an underlying mathematical model that could be directly tested. There were a few excellent applications of mathematics in biology, and the one that captured my was population genetics. To think that we could start to understand how evolution works with a combination of mathematics and observation seemed rather grand, but it simply rang true to me.

At the time the human genome was sequenced, it cost about a dollar to genotype one SNP in an individual. After the HapMap Project, chips for genotyping about one million SNPs in one individual cost about $500. Now there are chips for 250,000 SNP variants that cost $38. This ability to get genotypes cheaply greatly increased the effort to apply genome-wide association tests. The basic design for the GWA studies has been to score SNP genotypes in about one thousand diseased individuals and an equal number of healthy controls. By simple statistical association tests, many SNPs have been found to show a statistical overrepresentation in the diseased individuals. This research has identified many additional genes to study for their possible roles in diseases, yet it is clear that only a small portion of the genetic variation in disease resistance is captured this way. Who Is Related to Whom? GWA studies have greatly increased interest in trying to determine whether SNP genotyping can tell us who is related to whom and how far back we need to go to find a common ancestor. Modeling and statistical approaches have progressed rapidly, and now we have a very good ability to identify regions of the genome that are shared by common ancestry and regions that indicate nearest ancestry in Europe or Africa or Asia. These methods will not be able to pinpoint our ancestral past history precisely, but they are good enough to improve the accuracy of gene-mapping methods. In a recent study of variation in Europeans, the genetic variants clustered in a way that showed remarkable similarity to the map of Europe, suggesting consistency with a very simple model of isolation by distance. As this work on human ancestry has progressed, chimpanzee, gorilla, orangutan, gibbon, and macaque DNA have been sequenced, and we are able to ask which genes appear to evolve most rapidly in the human-specific lineage. It has proven remarkably difficult to pinpoint which genes give rise to traits that distinguish us from chimpanzees, but we are actually starting to get answers to these questions.

“You can learn a lot from me! There is more to this that meets the fly.� In addition, only last year, analysis of the genome of a Neanderthal made it absolutely clear that there was hybridization between modern humans and Neanderthal, and modern humans now have segments of Neanderthal genome in our sequence. How Recent Explosive Population Growth Affects Genetic Risk The most recent problem that Alon Keinan, Biological Statistics and Computational Biology, and I are tackling in human genetics concerns the influence of recent population growth on the accumulation of variants that increase disease risk. The human population has increased about a thousandfold in the last hundred generations, a staggering figure that puts us far out of the normal state of population genetic equilibrium. Basically, we have far too little variation compared to what the eventual equilibrium would be, if we were to maintain our current population size. This means that mutations are accumulating in the human population.

out-of-equilibrium state and the way these mutations are accumulating. All of these variants that arrived in the population in the last hundred generations are very rare in the population as a whole, so even if they inflate disease risk, very few individuals other than immediate family members will share the same risk alleles. This is why the study of population genetics in fast-growing populations is so timely— so that we can develop strategies for finding variants that affect disease risk. It also suggests that predicting disease risk from genetic variation will be much more difficult than we thought.

` clark-lab faculty/clark.cfm

Each human has about 100 new mutations, and when we sequence whole genomes today, we each differ by about three million bases out of three billion in the genome. Each new complete genome that is sequenced has between 50,000 and 150,000 variants that have never been seen before. This is expected, given our



Undergraduate Research


had, fulfill the American Dream that he had crossed the oceans for 25 years ago, with nothing in his pockets and a silent bride by his side, twisting and twisting the map of New York between her trembling fingers? He had built all of this up for them from the pitiless ground—the very lives they lived. And yet they were so unappreciative. He sighed, his heart heavy with pity. “Dad, you’re so proud!” his youngest son had cried out in adolescent throes of despair. “All you care about is your image, your reputation. How can you be so materialistic and go to church every Sunday? You never understand me!” Never understand? Who didn’t understand? — “Mr. Jeong Gets a Haircut”

*** Depending on the reader’s perspective, “Mr. Jeong Gets a Haircut” can either be a scathing satire or a heartbreaking story of an immigrant family. Ashley Woo ’12, undergraduate English major, does not tell her readers which it should be.

Through the Lens of an Undergraduate Writer For a fresh take on the immigrant generation gap story, the obstinate father speaks out in fiction. “We have to write support letters, Dad, it’s just part of going on mission trips.” Support letters? To other Korean adults in the congregation? Begging for money? No. Was it so his son could write to those clucking ajumas and harrumphing ajushis for alms that he had worked tirelessly for the past 25 years? When was the last time he had taken a day off for sickness? Had


he ever slept in late on a Saturday morning? Work, church, work, church, the cycle of his honest and hard weeks stretched behind him like a statue erected to the steadfastness of his character, the virtue of his sacrifice. Had he not worked himself to the bone, and then to the very marrow, so that his children could study at competitive schools, gain the advantage he had never

Mr. Jeong is a larger-than-life blend of typical Korean immigrant stereotypes— down to the model of his car and the blue-striped polo shirt. He has devoted everything to achieving his American Dream, but he must see it fulfilled through his children—his second-generation, Englishspeaking children, with their blaring music and meat-filled hamburgers. They bewilder him with crazy, idealistic desires to save the world, throwing away the opportunity that he desperately constructed for them. So he will make them do what he wants, because he knows what is best for them. It is a story that perhaps every child— Korean or not, immigrant or not—knows all too well. The second-place winner of the Arthur Lynn Andrews Prize for Fiction tells the story from the perspective of Mr. Jeong, that hard-working, selfless, obstinate, and overbearing father who loves his children, but will not listen to a word they say. As readers we instinctively side with the protagonist, but listening to Mr. Jeong

Undergraduate Research

is like listening to our own parents yell through the closed bedroom door; we find that we see from the perspective of his children, as well. “My intent while writing the story had been to get into the mind of a middle-aged Korean man and spend time there. That was all,” explained Woo, the American-born Korean writer. “I wanted readers to understand more, but not jump to a conclusion.”

touches readers. The logic whirling inside Mr. Jeong is hard to deny. He wanted an opportunity so badly and worked so hard to achieve it, but his children cannot assert themselves without falling into youthful, whiny pathos. On the other hand, they cannot understand where he is coming from because he never tells them, never communicates fully with them. We are driven speechless equally by Mr. Jeong looking for “some nice Korean boy” so

Depending on the reader’s perspective, “Mr. Jeong Gets a Haircut” can either be a scathing satire or a heartbreaking story of an immigrant family. Ashley Woo ’12, undergraduate English major, does not tell her readers which it should be. Woo said she found it strangely easy to tap into the perspective of an uncommunicative father—to speak his thoughts out loud. “It’s funny how it came to me,” she said, explaining how she was at Olin Library when the idea struck her. “It was one of those weird cases where I could very clearly visualize this project, if not Mr. Jeong, and I sat down and just wrote it in one go.” At the time, she was in English professor Maureen McCoy’s creative writing course, and she submitted the story for workshop. She received constructive criticism and comments from McCoy and other students in class. She remembers one particular Korean student who sent the story to his siblings back home because it reminded him so much of their father. Her older sister, an English alumna from Cornell, served as her editor, while her mother read the story over and corrected some details on Korean kimchi. Encouraged by McCoy to continue to work on the story, she revised and submitted “Mr. Jeong Gets a Haircut” to the competition. The rest, of course, is history. “I got some people saying that they hated Mr. Jeong,” Woo remembers. “I guess the outcry against Mr. Jeong was expected; I didn’t really write him as all sympathetic.” Yet something about the selfless tyrant

that his daughter does not take a white boy to prom, and by his son accusing his father of sending him to the Ivy League only to boast to other parents. With a simple plot and melodic language, Woo has portrayed both sides with admirable authenticity. The honesty in her writing allows the readers to feel for Mr. Jeong, despite everything. “I think to be honest about something, you need to have experienced it, to some extent,” she said. Woo’s own father was very flexible, she remarked, but watching other friends’ dads and men at church gave her that inner view, while still allowing her enough distance. “For a while I wanted to get away from being a Korean writer writing about Koreans,” she laughed. “But I think it’s what I know best and what I can write about the best.” Woo has tried being experimental in her writing, she admits, trying out wild styles and branching out from the Korean-American niche. Those efforts usually proved to be when readers misunderstood her stories, but they were also fun experiments that helped loosen her creative process. Woo is up in the air about her future. She is keeping all of her options open: she took the LSAT and is considering the possibility of an MFA program in creative writing somewhere down the line. Like all writers, however, she is pretty sure that even if she

doesn’t publish, she will always write. “Writing helps me to think better,” she said, “to observe more carefully. And it’s cathartic.” Joo Young Seo ’12 Joo Young Seo was an English major in the College of Arts and Sciences. She is attending law school at Columbia University.


Ashley Woo ’12

Working with Stuart Davis, English, Woo’s honors thesis was on the elusive Cornellian Vladimir Nabokov. She was awarded second place in the 2011 Department of English Barnes Shakespeare essay contest for her paper “Dismembering, Remembering, Speech” on Titus Andronicus. Woo is taking a gap year after graduation to experience the world and its wonders on a “grand or microcosmic scale,” as she puts it, before going to graduate school.

91 Photos in Undergraduate Research: Frank DiMeo unless otherwise noted

“Research helps me see, understand, and solve problems. It helps me think more critically and see beyond the surface to the next level.” Jiang and Henning used x-ray microtomography (microCT) imaging to observe the chick at different stages and construct a movie of its development in real time. MicroCT works like a human CT scan: x-rays create cross-sections of an object, which are then built into a 3-D virtual image. “Micro” refers to the high-resolution, micrometerscaled pixels of the resulting images, as well as the smaller physical size of the machine itself. This winning combination makes microCT a better and more cost-efficient alternative to traditional CT imaging for producing detailed images of smaller specimens.

Professor Jonathan Butcher, Alyssa Henning, and Michael Jiang

Agents for Seeing Contrast agents, which allow us to see inside living tissue, are not only crucial to imaging techniques, but also to extending the life of laboratory animals. Michael Jiang ’12 and Alyssa Henning ’11 were undergraduate researchers in the lab of Jonathan Butcher, Biomedical Engineering, tackling the world’s most common birth defect—the congenital heart defect (CHD). A congenital heart defect is any abnormality in the structure of the heart or heart vessels that arises during embryonic heart development. These abnormalities may be caused by genetic mutations or by environmental factors, such as prenatal infections, illnesses, or drugs. CHD affects approximately 9 of


1,000 births and is the leading birth defect–related cause of infant mortality. Jiang and Henning wanted to contribute to developing preventive measures and possibly treatments for human CHD by studying the hearts of embryonic chicks, which are quite similar to human hearts. By inducing heart defects in chicks during their development, researchers can better understand how defects form in humans and how they may be prevented.

Unfortunately, microCT on its own cannot effectively image soft tissue—contrast agents must be injected into the specimen to highlight soft tissue structures, such as the heart. These contrast agents shorten the lifespan of the embryonic chick, so that consistent images of a single chick’s development over time are difficult to achieve. In addition, the chicks need optimal heat and humidity conditions during the imaging process. Under the present unadjusted conditions, the chicks survive only about one day after imaging. Jiang and Henning decided to focus on finding the most effective and least toxic alternative to the usual contrast agent, osmium tetroxide. “Osmium tetroxide is the gold standard for imaging tissue, but it can only be used on dead tissue,” Henning explained. “We wanted to find a contrast agent that will keep the chicks alive. We’re developing a new protocol, since no one has imaged a live chick before.” Their study compared two specific agents: amnipaque and visipaque, which were injected on the fourth day of the embryos’ development. While similarly effective, visipaque was the least harmful to the

Undergraduate Research

Michael Jiang reflects, “Since I was able to graduate a year early, it was easy to choose to spend my fourth year pursuing a master of engineering degree in biomedical engineering at Cornell. I am happy to continue developing accessories for microCT as my master’s project.” Jiang plans to work in industry developing medical technologies. Alyssa Henning is attending Penn State University. She is in the Howard M. Salis Lab in synthetic biology. Her career goal is to “lead research projects that make biology easier to engineer.”

embryos, which survived for six days after injection on average. Jiang and Henning published their research, “Quantitative Three-Dimensional Imaging of Live Avian Embryonic Morphogenesis via MicroComputed Tomography” in Developmental Dynamics (July 2011). Henning commented that undergraduate research taught her how to approach complex challenges from multiple perspectives and that it’s okay if you don’t succeed the first time. “A lot of knowledge comes from the failed attempts. It took Mike and me two years before we finally got our microCT imaging protocol fully developed, Henning said. Although Jiang and Henning have different career goals, both agreed that undergraduate research provided them with valuable experience toward their future. Michael

Jiang received his bachelor degree in mechanical and aerospace engineering in 2011, one year early. He continued conducting research in Jonathan Butcher’s lab as a graduate student in biomedical engineering. Alyssa Henning received her bachelor’s degree in biological and environmental engineering in 2011 and went to work for Ginkgo Bioworks, a synthetic biology startup company in Boston, Massachusetts. Henning said, “I know that I want to get a PhD in biological engineering, biomedical engineering, or synthetic biology.” This decision, she said, is motivated in part by her undergraduate experience: “Research helps me see, understand, and solve problems. It helps me think more critically and see beyond the surface to the next level.”

Belinda Heyun Pang ’11 Belinda Pang is a PhD student in applied physics at the California Institute of Technology.


Michael Jiang: Alyssa Henning:



Center Spectrum


CNF Encompassing physical sciences, engineering, and life sciences with a strong interdisciplinary emphasis, the Cornell NanoScale Science and Technology Facility (CNF) supports a broad range of nanoscale science and technology projects.

ACSF The David R. Atkinson Center for a Sustainable Future (ACSF) is a bold initiative to focus the brightest minds from all disciplines—from the humanities to engineering—on creating new approaches, techniques, and technologies to advance solutions to some of society’s most complex challenges and pressing problems.

CNF is a national center that provides state-of-the-art resources and expert staff. CNF holds an annual meeting with research presentations, keynote speakers, poster sessions, and corporate soirees, offering an excellent opportunity for colleagues to learn about the exciting research by CNF users over the year.

The Institute for the Social Sciences (ISS) encourages collaborations among social scientists across disciplinary and institutional boundaries and engages the Cornell community in discussing current topics in the field.

To mark its 35th anniversary in 2012, CNF planned a special celebration for July, in lieu of its annual fall meeting. The guest speaker was William Brinkman, director of the Office of Science at the U.S. Department of Energy.

University Photography

// www.socialsciences.

Project leader Michael Jones-Correa “Panic Hits Home”

SHC is home to the Central New York Humanities Corridor, a crossdisciplinary research collaboration among humanists at Cornell, Syracuse University, and the University of Rochester, funded by the Andrew W. Mellon Foundation.

// sochum


A CNF Poster Session

Fellows will reflect on how the humanities converse with biological, ecological, economic, and technological approaches to risk. The year will feature a fellowship jointly sponsored by the Society and the Atkinson Center for a Sustainable Future.

Renate Ferro

University Photography

The newest interdisciplinary ISS theme project, Immigration: Settlement, Integration, and Membership, expands the theoretical frontier of immigration studies at Cornell. Opportunities for public participation during 2012 included a seminar series and workshops on labor immigration, the Dream Act, second-generation issues, new immigrant destinations, and the trend toward the criminalization of immigration.


Associate director Wendy Wolford

Cornell’s Society for the Humanities (SHC) brings distinguished visiting fellows, Cornell faculty, and graduate student fellows together each year to pursue research on an interdisciplinary focal theme. The focal theme for the 2012-2013 academic year will be RISK@Humanities.


Charles Harrington Photography

Wendy Wolford, Development Sociology, is the center’s new associate director of economic development. Wolford is an example of the interconnected work ACSF faculty fellows do in sustainability. Her research covers the political economy of development, social movements and resistance, agrarian societies, political ecology, land use, land reform, and critical ethnography, all with a regional concentration in Latin America, particularly Brazil. She is the author of a 2010 book on Brazil’s Movement of Rural Landless Workers, This Land is Ours Now: Social Mobilization and the Meanings of Land in Brazil.


Center Spectrum



Research at the Center on the Microenvironment and Metastasis (CMM) aims to unravel cancer’s complexity and understand the interaction of mechanical forces and chemical cues in cancer metastasis. The center pursues experimental and theoretical approaches derived from the physical sciences in order to address major questions and barriers in understanding and treating cancer.

The Cornell Center for Materials Research (CCMR) is a springboard for innovative materials science and engineering research. CCMR Shared Facilities offer world-class materials characterization, analysis, and processing equipment. Electron and optical microscopy, spectroscopy and electronic measurements, and surface analysis and characterization are all available at the center.


Frank DiMeo

The U.S. Department of Energy’s Office of Basic Energy Sciences supports the core research of EMC2. With support from New York State and others, the center works with industrial partners, including General Motors, Subaru, and Primet Precision Materials, to adopt novel materials into advanced energy technologies.

CCMR Microscopy Facilities host an annual competition for the best image produced using an electron microscope. In the 2011 competition, Pinshane Huang (advisor David Muller, Applied and Engineering Physics) won the Overall Award for Visual Impact, using the FEI Spirit to produce polycrystalline graphene that resembles an atomic patchwork quilt.


Project one, led by Claudia FischbahTeschl and Vivek Mittal, integrates physical sciences and cancer biology approaches to enhance understanding of the mechanisms behind tumor vascularization. Project two, led by Cynthia Reinhart-King and Paraskevi Giannakakou, focuses on how chemical and mechanical forces in tumors enable and increase cell migration during metastasis. Project three, led by Michael King and David Nanus, focuses on understanding the fundamental physical mechanisms of circulating tumor cell adhesion to inflamed endothelium under flow.


The Energy Materials Center at Cornell (EMC2) advances the science of energy conversion and storage. Faculty from chemistry, materials science, chemical engineering, and physics work with postdocs and student researchers to grow materials for new generations of batteries and fuel cells. The center uses new techniques, such as in-situ liquid TEM, to analyze complex oxides.

First-place winners include Byungki Jung, (advisor Michael Thompson, Materials Science and Engineering); Kaifu Bian (advisor Tobias Hanrath, Chemical Engineering); Ye Zhu (advisor David Muller, Applied and Engineering Physics); and Amy Blakeley (advisor Lara Estroff, Materials Science and Engineering). Recently announced 2012 winners can be found at contestimages/winners12.

// facilities




Project leader Cynthia Reinhart-King Director Héctor Abruña Graphene Strain-induced ferroelectricity and strain relocation


CLASSE Looks Ahead A Conversation with Ritchie Paterson


In 2006, Connecting with Cornell highlighted the Cornell Laboratory for Accelerator-based Sciences and Education (CLASSE), showcasing Cornell’s wide-ranging expertise in particle physics, accelerator technologies, and x-ray sciences. Will you describe the activities of CLASSE? CLASSE is tremendous! CLASSE encompasses research and education projects involving dozens of faculty, hundreds of staff and undergraduate and graduate students, and thousands of collaborators and CHESS facility users from around the world. CLASSE continues a stellar trajectory in 2012, as the center welcomes a new director, Ritchie Patterson, whose Cornell career spans undergraduate student, CHESS operator, and Department of Physics faculty member and chair. Patterson’s research centers on experimental particle physics using the Large Hadron Collider near Geneva, Switzerland. Why the move from chairperson of the Department of Physics to director of CLASSE, taking the reins from retiring director Maury Tigner?

Right now, we have more than 25 externally funded ongoing projects in CLASSE. I invite readers to follow the news of CLASSE scientists. These innovators are probing symmetries that guide the universe, producing and sustaining beams of incredibly closely packed particles, developing energy-efficient ways to accelerate particles, and pushing the fundamental limits of light sources and other accelerators.


PAT T E R S O N : CLASSE is pushing the boundaries in the physics of beams and accelerators, in exploring the cosmos, and in developing x-ray tools that enable new discoveries and inventions in the biomedical, materials, and environmental sciences. And students are deeply involved in all of this outstanding research. It’s a fantastic privilege for me to be a part of the groundbreaking science at CLASSE and to continue a strong commitment to training the next generation of scientists.


Less publicized but vitally important are the hundreds of completed and ongoing PhD thesis projects at CLASSE. Cornell is one the few universities in the country that is preparing accelerator physicists equipped to build x-ray sources crucial to scientific progress in many fields. One CLASSE project is an ERL prototype. What is the ERL, and what will it bring to the Cornell community? The Energy Recovery Linac will be a firstof-its-kind high-energy x-ray source, using accelerator technology invented and being perfected here at Cornell.

Maury Tigner proposed in 1965 that highly tuned resonant cavities could be used both to accelerate and decelerate charged particles. During deceleration, the energy of the particles is recovered and recycled. Realizing this elegant idea was beyond existing technical capabilities, 10 years ago CLASSE started an NSF-funded R&D project to build a prototype source. As the prototype nears its goals, we’ve developed a technical design plan for a full-scale facility. Last summer we hosted six international workshops exploring how an ERL opens new avenues for research that uses continuous-duty coherent beams of ultrashort x-ray pulses. This helped us to clarify a science case for such a facility. Cornell faculty and students were key participants looking forward to a powerful new tool on campus for investigating all types of materials, from airplane wings to cell membranes, and from pollution in plant tissue to matter under earth-core pressures. Recent ERL press releases tout some major technical milestones. What do these milestones predict about the future of a new x-ray light source at Cornell? CLASSE is building accelerator components that were only a dream when the R&D project began. These components are now exceeding world performance records. While we have more work to do, our technical accomplishments show that an ERL is within our reach. The milestones demonstrate that CLASSE has innovative, world-leading capabilities— and talent—and that Cornell’s Ithaca campus, with outstanding research and education, is the best place in the world to build a first-of-a-kind facility. Given the uncertainties we hear about funding for science, are you confident that CLASSE can continue making progress in the coming years?

Center Spectrum

Success in funding follows a solid track record and vision. Cornell has been a leader in accelerator-based sciences for over four decades, consistently building a trail of innovative and daring “firsts.” As long as our faculty and students generate and substantiate new ideas, I am confident that CLASSE will continue to grow. Will you comment on the educational projects and initiatives that CLASSE supports? It’s essential to the health of our nation to get citizens, and especially young students,

engaged and excited about science, technology, engineering, and math—the STEM fields. And I can think of no better place for young people to discover that excitement than here at CLASSE and Cornell. CLASSE has two full-time science educators, Lora Hine and Erik Herman, designing and delivering education and outreach programs. Each year, thousands of elementary and high school students participate in our programs at the lab and in the community.

more girls and young women into science, so I’ve just committed to hosting the Northeast Conference for Undergraduate Women in Physics, which will bring over a hundred promising young scientists to campus in early 2013.


We also support summer research experiences for college students from across the country. I am always excited about getting




To see the smallest details ERL MILESTONES Future technology—from computer chips to pharmaceuticals— depends on seeing the smallest features. The Cornell ERL will enable scientists to visualize motion at the molecular level in cells; study the smallest details of metals, ceramics, polymers, and other materials; and follow chemical reactions in ways never before possible. With these capabilities, ERL x-ray beams will help scientists develop more efficient engines, batteries, photovoltaics, and fuel cells; analyze pollutants and environmental toxins; create energy-efficient materials and

technologies; and much more. The ERL will be a new type of continuous-duty, short-pulse x-ray source, using ultracompact electron bunches and a superconducting linear accelerator (linac) that will accelerate and recycle particle energy. To prove that such a source is possible, CLASSE has been designing and building unique photoinjector and superconducting cavities. This NSF-funded prototyping project has been achieving milestones— and, in some cases, extending world records—every day!


Milestone 1: A continuous-duty current of 50 milliamperes from the laser-driven photocathode electron gun sets a new world record and exceeds the levels needed by one of the ERL operating modes. The full-scale ERL operating mode of 100 mA is within sight. Milestone 2: The brightness of the prototype gun, produced by the core of the electron bunches (the central two-thirds of the electrons), already equals what is needed for a full ERL. Better values are expected when the

injector voltage is increased. This super-bright core was unexpected at the start of the project. The discovery could dramatically advance the capabilities of an ERL over existing sources. Milestone 3: Superconducting accelerating cavities need to be extraordinarily efficient for an ERL linac to recover and reuse electron beam energy. The first ERL prototype accelerating cavity achieved an efficiency surpassing ERL requirements. If a church bell or chime were this efficient, it would ring for a whole year after a single strike!

99 Photos in this article: CHESS; Frank DiMeo


$ $

Research Newly Funded

$ 101


Research Newly Funded

Selected Sponsored Research Awards Andrew H. Bass

Todd R. Evans

Melissa A. Hines

Catherine Lord

Neurobiology and Behavior Neural and Hormonal Mechanisms of Vocal Communication National Science Foundation $790,712

Surgery, WCMC A Molecular Pathway Controlling Cardiomyocyte Specification National Heart, Lung, and Blood Institute $2,112,500

Chemistry and Chemical Biology Cornell Center for Materials Research National Science Foundation $18,359,999

Psychiatry/Westchester, WCMC 2/2-Effects of Parent-Implemented Intervention for Toddlers with Autism Spectrum (ESI) National Institute of Mental Health $1,416,017

Julien V. Koschmann and Hirokazu Miyazaki East Asia Program The Bamboo Texts of the Guodian: a Study and Complete Translation, Volumes One and Two Chiang Ching-Kuo Foundation $5,000

Kenneth P. Birman Computer Science Platform for High-Assurance Cloud Computing (Hac2) Department of Defense $4,949,804

Cynthia Ruth Farina, Claire T. Cardie, and Daniel R. Cosley Law School/Computer Science/ Information Science Improving Citizen Participation in Rulemaking: Integrating Human and Computational Facilitation of Online Deliberation in Complex Government Policymaking National Science Foundation $750,000

Dan Luo Biological and Environmental Engineering DNA Nanotechnology-Based Detection for Multiplexed On-Site Diagnosis of Plant Pathogens USDA $499,881

Michael I. Kotlikoff Biomedical Sciences Genotyping Platform Pfizer $500,000

Tanzeem K. Choudhury Information Science Enabling Community-Scale Modeling of Human Behavior and Its Application to Healthcare National Science Foundation $424,125

Paulette Clancy and William Robert Dichtel Chemical and Biomolecular Engineering/ Chemistry and Chemical Biology Ultimate Electronic Device Scaling Using Structurally Precise Graphene Nanoribbons National Science Foundation $1,200,000

Chris Fromme Molecular Biology and Genetics/ Weill Institute Regulation of ARF GTPASE Activation at the Trans-Golgi Network National Institute of General Medical Sciences $1,464,678

Molecular Biology and Genetics Population Genetic Inferences from Dense Genotype Data National Human Genome Research Institute $2,304,667

Joanne Difede Psychiatry, WCMC Enhancing Exposure Therapy for PTSD: Virtual Reality and Imaginal Exposure with Cognitive Enhancer Department of Defense $11,000,000

Geraldine K. Gay

Electrical and Computer Engineering Wafer-Scale Tunable X-Ray Sources for Phase Contrast CT-Scanning Department of Defense $2,464,546

Biomedical Engineering/Weill Institute Systematic Analysis of the Functional Relevance of Nuclear Structure and Mechanics in Breast Cancer Progression Department of Defense $667,112

Molecular Biology and Genetics Genetic Regulation of Phospholipid Synthesis in Yeast National Institutes of Health $2,045,196

Susan R. McCouch Plant Breeding and Genetics Enhancing Utilization of Crop Wild Relatives: Capturing Genetic Value from Ancestral Populations of Wild Rice Global Crop Diversity Trust $191,017

Alyosha Molnar

Communications Collaborative Research: Improving Online Political Deliberation with Computational Supports for Frame Reflection National Science Foundation $483,019

Susan A. Henry, Manuel Aregullin, Yu-Fang Chang, Maria L. Gaspar, and Stephen A. Jesch

Horticulture/Applied Economics and Management/Food Science Northern Grapes: Integrating Viticulture, Winemaking, and Marketing of New Cold-Hardy Cultivars Supporting New And Growing Rural Wineries USDA $2,511,333

Amit Lal

Jan Lammerding, Patricia Davidson, and Celine M. Denais

Andrew G. Clark

Timothy E. Martinson, Miguel Gomez, Kevin A. Iungerman, and Anna K. Mansfield

Electrical and Computer Engineering Integrated Systems for Light Field Capture and Analysis National Science Foundation $400,000

Ruth E. Ley and Andrew G. Clark Microbiology/Molecular Biology and Genetics Genetic Transmission of Components of the Human Gut Microbiome National Institutes of Health $1,728,828

John P. Moore Microbiology and Immunology, WCMC Structure and Immunogenicity of Cleaved, Stabilized HIV-1 Envelope Trimers National Institute of Allergy and Infectious Diseases $7,753,805

$ 102 Photos on this page: Frank DiMeo; Lindsay France/CU; Jason Koski/CU; University Photography; Provided

Jenny E. Sabin

eBraid City Plan Jenny Sabin, Architecture


Jenny E. Sabin




Research Newly Funded

Insight into how cells modify their immediate extracellular matrix (ECM) microenvironment with minimal energy and maximal effect will lead to the biomimetic design and engineering of highly aesthetic, passive materials, along with sensors and images that will be integrated into responsive building skins at the architectural scale. Jenny E. Sabin Architecture Energy Minimization via Multi-Scalar Architectures: From Cell Contractility to Sensing Materials to Adaptive Building Skins University of Pennsylvania $300,972

Terence L. Robinson, Lailiang Cheng, Alison M. Demarree, Stephen A. Hoying, and Bradley J. Rickard

Ling Qi Nutritional Sciences Dissecting the Role of Unfolded Protein Response in Alcoholic Liver Disease National Institute on Alcohol Abuse and Alcoholism $404,250

Horticulture/Applied Economics and Management Enhanced Economic Sustainability of Small Farms through the Production of Stone Fruits Using Mesoclimatic Modification Technologies USDA $499,917

Mark E. Sorrells and Michael H. Davis Plant Breeding and Genetics Value-Added Grains for Local and Regional Food Systems USDA $2,356,999

Michael O. Thompson and Dieter G. Ast Materials Science Engineering Substrate Interactions with Ingazno Thin Film Transistors Corning $916,655

Bradley J. Rickard Applied Economics and Management Optimal Licensing of an Agricultural Innovation USDA $348,676

Richard Robinson Materials Science Engineering Nanoscale Phonon Spectrometer to Quantitatively Characterize Low-Dimensional Heat Transfer National Science Foundation $600,000

David Walter Wolfe, Antonio Miguel R. Bento, Stephen D. DeGloria, Jeffrey Melkonian, and Harold Van Es Horticulture/Applied Economics and Management/Crop and Soil Sciences New Tools and Incentives for Carbon, Nitrogen, and Greenhouse Gas Accounting and Management in Corn Cropping Systems USDA $4,705,170

Jocelyn Rose and Larry P. Walker

Chris Xu and Frank W. Wise

Plant Biology/Biological and Environmental Engineering Defining Determinants and Dynamics and Cellulose Microfibril Biosynthesis, Assembly, and Degradation Department of Energy $2,005,372

Applied and Engineering Physics

David G. Russell Microbiology and Immunology

Derek H. Warner Civil and Environmental Engineering Characterizing Stiffness Degradation in High-Performance Welded Aluminum Structures Department of Defense $152,000

Technology Development for In Vivo Deep Tissue Imaging National Institute of Biomedical Imaging and Bioengineering $1,638,269

The Role of Granuloma in M. Tuberculosis Infection National Heart, Lung, and Blood Institute $1,740,955

$ 105 Photos on this page: Frank DiMeo; Lindsay France/CU; Jason Koski/CU; University Photography




n 107


Seeing DNA and Beyond: Biotechnology Resource Center (BRC)

The BRC makes high-end biotechnologies available to all researchers, provides a wide range of instruments and services, and supports research and education.




Busting through the Unseeable What do you do if you want to see ghosts? Call Ghostbusters! But what if you want to see infinitesimally small structures in a plant or animal? Call the BRC, the Cornell University Biotechnology Resource Center (formerly called the CLC, the Life Sciences Core Laboratories Center). Illuminating genetic variation, protein modifications, cells and tissue structures, and many other fundamental characteristics of life, the BRC’s cutting-edge biotechnologies enable exciting and productive ways of seeing. These technologies are critical to understanding development, disease, and interactions between organisms and the environment. The BRC’s laboratories offer an exceptional opportunity for creative applications with the potential for breakthrough discoveries. The center is composed of six core facility laboratories focused on genomics, proteomics, imaging, bioinformatics, bio-IT, and advanced technology assessment. The BRC’s array of state-of-the-art tools and technologies are available to researchers at Cornell and other institutions. Located on the Cornell campus in Ithaca, the BRC incorporates fee-forservice research, technology testing and development, and educational components. Investigators spanning many life science disciplines and other fields utilize the BRC’s technologies and expertise. In the past year, 628 investigator groups used the facilities, including researchers from 79 departments in seven Cornell colleges and from 129 institutions in 38 states and 15 other countries. The center expects many new researchers and research projects in the coming years. Reviewing protocols and planning experiments carefully to leverage the technologies is critical, given the broad range of capabilities and rapidly developing applications of the instruments. Accordingly, the BRC offers coordinated project consultations with the directors and staff of all the core services during the design, data production, and analysis phases of investigators’ research projects. Reading the Letters in the Book of Life Reading the sequence of the basic components of DNA is like reading the letters in

the Book of Life. The BRC Genomics Facility’s cutting-edge DNA sequencing technologies provide instruments and services that enable investigators to view genes and genomes in very fine detail. The facility supports a wide range of research, including studies of the composition and organization of whole genomes, genetic variation between individuals and between species, the role of genetic mutations in development and disease, and analyses of gene expression and regulation.

The instruments in the proteomics core enable large numbers of proteins from complex biological samples to be detected and analyzed with high resolution and accuracy. Recent research projects include studies of regulatory networks and the assembly of protein complexes and signal pathways in bacteria, plants, and animals, as well as projects on the role of protein interactions and modifications in infertility, diabetes, and cancer.

The BRC Imaging Facility provides many ways of seeing, from real-time movies of the heart beating in a living mouse, to images of paperthin structures lying underneath the skin of a tomato, to detailed three-dimensional digital reconstructions of the exoskeleton of trilobites embedded in fossils that are hundreds of millions of years old. The core’s resources include sequencing, microarrays, and genotyping technologies. New DNA sequencing technologies are emerging at a dizzying rate, and the genomics core has been actively acquiring and implementing the latest sequencing instrumentation. The new generation of sequencing technologies enables many new research applications, including fast and cost-effective whole genome and targeted region sequencing of humans and other organisms. In the past year, the genomics core analyzed about 500,000 samples from 500 investigator groups and generated about 10 trillion bases of DNA sequence. Seeing the Building Blocks of Life The BRC Proteomics and Mass Spectrometry Facility provides tools for identifying and characterizing proteins and other small molecules that are the building blocks of life. The facility has mass spectrometers and other analytical tools for identifying and quantifying proteins and small-molecule metabolites, for detecting chemical modifications that can affect protein function, and for quantitative proteomics.

A Picture Paints a Thousand Words The BRC Imaging Facility provides many ways of seeing, from real-time movies of the heart beating in a living mouse, to images of paper-thin structures lying underneath the skin of a tomato, to detailed three-dimensional digital reconstructions of the exoskeleton of trilobites embedded in fossils that are hundreds of millions of years old. The imaging core has tools for directly viewing structures ranging in scale from subcellular units like chromosomes and mitochondria, to whole cells and tissues and organs, to whole organisms. Image platforms include • transmitted and fluorescence light microscopy • confocal and stereo microscopy • luminescence and fluorescence imaging • multiphoton microscopy • flow cytometry analysis • micro and nanoscale x-ray microscope computed tomography (CT) • high-resolution ultrasound Imaging instruments, like the micro and nano CT, are also used to support research in fields outside of the life sciences—for

n 110 Photos in Outreach: Biotechnology Resource Center and Frank DiMeo unless otherwise noted

3D visualization of blood vessels within a dog heart, generated by a micro-scale computed tomography (CT) instrument (GE eXplore CT120) in the BRC Imaging Facility (M. Riccio, F. Fenton, R. Gilmour, Jr.)

3D visualization of the internal structure of a Gar fish, generated by a micro-scale computed tomography (CT) instrument (GE eXplore CT120) in the BRC Imaging Facility (A. McCune, M. Riccio)

The core staff has developed and supports research databases for national and international basic and clinical research projects. example, in studies that need to see the inner structure of nanofabricated devices. Supporting a Fire Hose of Data The BRC Bioinformatics Facility (also called the Computational Biology Service Unit) supports analysis of the massive amount of data generated by the BRC core facilities. It provides large-scale computational resources, a diverse array of software analysis tools, and extensive expertise in their use. The BRC bioinformatics core is a Microsoft Institute for High-Performance Computing—one of only 10 worldwide. The bioinformatics core has three computational clusters with a combined total of more than 1,200 processers and 500 nodes to support data analysis. This infrastructure supports projects in genomics, quantitative genetics, proteomics, and structural biology.

n 112

One particular area of expertise is the development and management of large-scale research project databases that integrate many types of basic research data, such as genomics and proteomics information joined with phenotyping data.

and complex new era of life science research, so the facility gives high priority to training and education workshops.

The core staff has developed and supports research databases for national and international basic and clinical research projects. The facility also created and supports BioHPC, a high performance computing suite of more than 50 life sciences software applications, and maintains the BioHPC Laboratory for biologists who want to learn the Linux operating system and how to do large-scale genomic data analysis.

At the Heart of the BRC The BRC Bio-Information Technology (Bio-IT) Facility creates and maintains the information technology infrastructure that enables the operations of all the BRC’s cores. This vital infrastructure includes a Laboratory Information Management System (LIMS) for tracking and processing sample submissions, instrument use, data storage, and data transfer to investigators. The Bio-IT Facility also provides high-end IT support for life sciences research groups and academic departments at Cornell.

The data deluge witnessed at the BRC Bioinformatics Facility—a veritable fire hose of information—promises an exciting

Kicking the Tires As new sophisticated analysis technologies are developed and become commercially



BRC Core Facility Directors

Genomics Peter Schweitzer Proteomics and Mass Spectrometry Sheng Zhang Imaging Rebecca Williams and Mark Riccio

available, their implementation can be a major headache. Researchers want access to cutting-edge instrumentation, but all too often the latest exciting piece of technology is far from “plug and play.” Many wrinkles need to be shaken out. The Advanced Technology Assessment (ATA) Facility works closely with the other BRC core facilities and with Cornell faculty to identify the emerging technologies needed to keep scientists at the cutting edge of research, to bring these instruments to campus for testing, and to acquire funding for their purchase. Once the new instruments are installed, the ATA core helps to “kick the tires,” optimizing existing methods and developing novel applications for these platforms. The goal of this core is to

expedite early-phase optimization and to get to the point of high throughput data generation as smoothly and rapidly as possible. Championing Research and Education Sharing the Expertise. Scientists from other national and international institutions visit the BRC to learn about new life sciences shared resource technologies. These scientists often hope to start a biotechnology center at their home institution. Scientists and graduate students visiting the BRC facilities for short tours and extended stays have come from as far as Niger, Brazil, Sri Lanka, India, Ghana, Indonesia, Spain, and China.

Bioinformatics Facility Jaroslaw Pillardy and Qi Sun Bio-IT Facility James VanEe Advanced Technology Assessment George Grills

Jocelyn Rose Director, Institute of Biotechnology Linda Carr Executive Director, Institute of Biotechnology George Grills Director, Operations




The core center’s broad range of high-technology facilities, menu of instruments and services, and support of multidisciplinary research plays an important role in the recruitment and retention of faculty.

Multidisciplinary Collaboration and Faculty Recruitment. The BRC facilitates collaborations between departmental disciplines and between institutions. The core center’s broad range of high-technology facilities, menu of instruments and services, and support of multidisciplinary research also plays an important role in the recruitment and retention of faculty. Advancing the Professional. To facilitate broad and effective use of their high-end instruments, all BRC core facilities organize regular educational workshops and seminars. The BRC bioinformatics core, for example, hosts a genomics bioinformatics educational workshop series that covers the practical aspects of using computational biology applications in HPC environments. Faculty, postdocs, students, and laboratory research staff attend these educational workshops. The ATA core, as another example, sponsors a university-wide Life Sciences Advanced Technology Seminar series that has presentations and discussions on emerging biotechnologies. To inform the Cornell community about the BRC’s facilities and to promote regional sharing of these research resources and services, lectures are posted online and are presented at local, regional, national, and international meetings. BRC core directors also give presentations in undergraduate and graduate courses and at departmental seminars.

n 115













Data generated in the BRC Imaging Facility: [left to right starting with top row] (a,b) ovary development in fruit flies (Erin Kelleher, D. Barbash lab); (c) cells growing in a microfabricated channel (Mandy Esch, T. Stokol lab); (d) green fluorescent protein in epidermal cells of a plant leaf;

n 116

(e) endoplasmic reticulum in plant epidermal cells (biology class lab, M. Hanson); (f) cells and subcellular structures; (g) dividing cells in a fruit fly embryo (Byron Williams, M. Goldberg lab); (h) cells and subcellular structures; (i) cells with nuclei (blue), plasma membranes (green), and

mitochondria (red) (R. Williams and W. Zipfel); (j) algae and zooplankton; (k) ovary development in fruit fly (Shamoni Maheshwari, D. Barbash lab); (l) dispersion of oils in water for food production (Wen Chyan Tsai, S. Rizvi lab)

The BRC helps support outreach programs for teaching, training, and learning at the high school, undergraduate, and graduate levels, including a training program for high school biology teachers. A Technology Fest. The BRC organizes the annual Cornell Life Sciences Research Resources Expo, which is held on the Ithaca and New York City campuses. This annual technology fest includes a public seminar and poster session, with presentations by life science cores from all of the Cornell campuses. NERDS. The BRC sponsors the annual Northeast Regional Life Sciences Core Directors (NERDS) meeting, which is a forum for topics of interest to biotechnology cores. This meeting has grown enormously popular over the last few years: 158 core directors from 60 institutions and 21 states attended the NERDS 2011 meeting, which was held on the Ithaca campus. Reaching Out to Educate. The BRC helps support outreach programs for teaching, training, and learning at the high school, undergraduate, and graduate levels,

including a training program for high school biology teachers. The center is also involved in programs that promote and broaden the participation of women and underrepresented minorities in science. The BRC Organization The BRC is administered by the Cornell University Institute of Biotechnology and is part of a New York State Center for Advanced Technology (CAT) called the Center for Life Science Enterprise. Each BRC core has a facility director and a faculty advisory board that provides scientific guidance.

George Grills Director of Operations Biotechnology Resource Center Jocelyn Rose Director, Institute of Biotechnology Associate Professor of Plant Biology


The Cornell University Biotechnology Resource Center provides innovative ways of seeing fundamental characteristics of living organisms. Looking forward, the BRC will be key to keeping Cornell at the forefront of a wide range of dynamic and rapidly evolving life sciences research.

n 117


ue u e

Technology Transfer

u e


Laser Focused Applying Physics According to Frank Wise, Applied and Engineering Physics, “An applied physics success story is, first and foremost, good physics— then physics good enough for products people will pay for.” Over the past two decades, Wise has played a critical role in both the advancement and commercialization of femtosecond lasers. Femtosecond lasers are remarkable tools. They make pulses of light that last about 100 femtoseconds, with a femtosecond equal to a millionth of a billionth of a second. Femtosecond lasers are able to cut

u e


almost any material without heating it—if the material is living tissue, without causing any pain. These amazing devices have applications in biological sciences, medicine, advanced energy, and manufacturing. Wise’s research group has been a world leader in recent efforts to make cheaper,

more robust, and more powerful fiber femtosecond lasers, which promise to expand the adoption and use of femtosecond lasers dramatically. On the Way to a Laser Focused Obsession Wise’s physics career got off to a somewhat late start. It wasn’t until he went to Princeton, where he graduated with a BS degree, that he found the intellectually stimulating environment of science. He earned a master’s degree in electrical engineering from the University of California–Berkeley, and then went to work at Bell Labs on integrated circuit development. After two years at Bell Labs, Wise arrived in Ithaca in 1984 to begin a Cornell PhD in physics with Chung Tang. He has been at Cornell ever since.

u e

While Wise was pursuing his PhD, scientists had begun to develop femtosecond lasers, but they were not yet commercially available. Femtosecond dye lasers were difficult to build. Tang’s group was one of a few groups worldwide working on femtosecond lasers. Soon after Wise joined Tang’s group, he and another student were able to apply a critical advance in the understanding of femtosecond pulse formation to a femtosecond dye laser they built. At the time, their achievement was just a tool for Wise’s research in semiconductor materials, but it soon became a second leg of his research. Commercializing a First One day on campus, Wise ran into Bill Clark, a colleague from his time at Bell Labs. Clark was now working at Newport Corporation. When he learned that Wise had made his own dye laser, Clark hired him to design and build a mode-locked dye laser kit with an instruction book and video. Clark demonstrated the instrument at a conference, where it was well received. Newport decided, however, that it did not want to pursue the product. So Clark quit Newport and started his own company to make and sell the dye laser kit Wise had designed. This was the first commercial femtosecond laser. Femtosecond Laser Meets Two-Photon Microscopy After completing his doctoral program in 1988, Wise accepted a faculty position in the engineering college. His research group began building more femtosecond lasers for studies of ultrafast phenomena. Soon another faculty member on campus, Watt Webb, Applied and Engineering Physics, and his postdoc, Winfried Denk, conceived of two-photon excitation microscopy, and they needed some way to test it. Two-photon excitation microscopy is a fluorescence imaging technique that allows imaging as much as one millimeter beneath the surface of living tissue. The technique works by focusing laser pulses on a very small volume of tissue, causing molecules within that volume to fluoresce. A very high-intensity pulse is needed—but one that will not damage the tissue with high average power. It was just the job for a femtosecond laser.

Technology Transfer

Webb brought his microscope to Wise’s lab for the first demonstration of two-photon excitation microscopy. Two-photon excitation microscopy ultimately became one of Cornell’s biggest technology transfer successes, and related products are now being sold worldwide under a Cornell license by Zeiss. The Serendipitous Advent of the Ti-Sapphire Laser Around this time, researchers at Saint Andrew’s University in Scotland accidentally made a discovery. They had been working on a new solid-state laser, based on titanium ions (Ti) in a sapphire crystal. The laser they built did not work as intended, but in a stroke of serendipity, their laser emitted extremely short and powerful pulses. These new Ti-sapphire lasers were the first practical femtosecond lasers, a discovery that made possible the commercialization of two-photon microscopes, when femtosecond dye lasers proved unsuitable. Naturally, Wise’s group built one. He again helped Bill Clark’s company by showing him how to build a Ti-sapphire femtosecond laser. As a result of this collaboration, Clark’s company had the first Ti-sapphire laser on the market, and within a few years, had sold more than 250 to labs around the world. Making a Better Laser Over time, lasers became a captivating area of research for Wise, apart from his study of nanostructures for physics, light emitters, and solar cells. After working with dye and Ti-sapphire lasers for a number of years, Wise decided around 2000 to try to make short-pulse fiber lasers. Fiber lasers have an optical cavity made from optical fiber, and instead of bouncing back and forth between two mirrors, light travels around a loop of fiber multiple times until it builds up enough energy to escape as a pulse. Other research groups had been trying to build femtosecond fiber lasers, but the resulting instruments were unstable or produced pulses with very low energy. With his experience in short-pulse lasers, Wise thought his group stood a good chance. As he said, “When we go into a new area, we give it about a year, and if we don’t have any new ideas by then, we drop it.”

Laser Notes from Frank Wise

Ultrafast science and technology are dominated by solid-state lasers, such as Ti-sapphire. These are great laboratory tools. They are about one meter long and half a meter wide and cost $120,000. Light bounces back and forth between precisely aligned mirrors, which is the case in most lasers. When we first get an idea for a new fiber laser, it looks like the instrument pictured above. You can see that it is part fiber and part discrete components that allow us to adjust things and see what happens. These lasers are experiments. Once we understand things, we make them entirely of fiber. A fiber laser that will soon compete with solid-state lasers is shown below. The light pulses just go around in the fiber, so there is nothing that can become misaligned. The box is five by four inches, and a power supply about the same size will be needed. Such a laser will be much cheaper and easier to operate than the solid-state laser, so it will find broad application outside research labs.

u e

121 Photos in Technology Transfer: Frank DiMeo and Wise Lab unless otherwise noted

? r se a L nd co se to m e F a is How Powerful The ultrashort pulse of light emitted by a femtosecond laser beam is so powerful that it directly breaks chemical bonds in the material being cut. The pulse is too brief to transfer heat or shock to the material, allowing femtosecond lasers to drill holes or make cuts in almost any material with virtually no damage to the surrounding area. In contrast, a laser with a longer pulse melts and boils away surrounding material as it cuts.

u e

More Inventions Wise and his students did have new ideas and good physics, and as a result, began to progress and create patentable intellectual property. Wise gives his students rein to explore and solve problems, and his students have repaid his trust. Great students, Wise says, can tolerate ill-defined thesis projects. He credits great students—Ömer Ilday, Andy Chong, and Will Renninger—with being instrumental in making the advances that led to practical, high-performance femtosecond fiber lasers. Ilday helped to invent the self-similar laser, while Chong and Renninger worked on developing the all-normal-dispersion (ANDi) laser. Laser Challenges The biggest problem with fiber lasers is the nonlinearity that always limits the propagation of short-pulses within a fiber cavity. Strong pulses change the properties of the fiber through which the pulse is travelling, so that as the pulse gets stronger, the fiber behaves differently. Builders of early fiber lasers tried to avoid nonlinear effects by stretching the pulse to prevent it from getting intense enough to damage the fiber’s optical properties. It proved impossible to avoid all nonlinearity, especially as researchers tried to make the lasers more powerful, with shorter pulses. Wise’s group wanted to develop fiber lasers that performed as well as the Ti-sapphire laser, the solid-state laser standard. The next step was to make pulses that tolerate nonlinearity. The Self-Similar Laser The efforts of Wise’s group to develop pulses tolerant of nonlinearity led to an exciting advance: the self-similar laser. This laser is constructed so that the pulse will propagate through the laser self-similarly. To propagate self-similarly means that, while the amplitude and phase of the pulse may vary as it propagates around the optical cavity, the shape of the pulse does not. Because the self-similar pulse can tolerate a great deal of nonlinearity, the performance of the self-similar laser is significantly better than earlier fiber lasers—although it could still be improved. The self-similar laser is under commercialization by a European company that makes lasers for machining applications. ANDi: Another Big Advance The biggest advance came when Wise

Technology Transfer

decided to create laser pulses that don’t just tolerate nonlinearity, but actually thrive on it. The result was the ANDi laser. The key to the ANDi laser is a filter. Nonlinearity tends to broaden the spectrum of wavelengths in a pulse; the filter cuts off the edges of the spectrum and shapes the pulse in time, so that nonlinearity is actually exploited to make the pulses. This simple process allows lasers to be built without complicated components, such as prisms, diffraction gratings, or microstructure fibers that were needed in all previous femtosecond lasers. The ANDi laser produces pulses with the highest possible stable pulse energy. Even the early ANDi lasers offered performance comparable to that of solidstate lasers, with the major benefits of fiber: reduced cost, alignment-free operation, increased stability, and compact construction. The most recent ANDi lasers from Wise’s lab are able to produce short pulses at considerably higher levels of power than solid-state lasers. The ANDi laser is easy to manufacture, reliable, and efficient. It is the first all-fiber laser that can directly replace a Ti-sapphire laser in most applications. Commercializing ANDi Multiple companies are developing products based on the ANDi laser and are negotiating with Cornell for licenses. The Cornell Center for Technology Enterprise and Commercialization (CCTEC) has deemed the fiber laser technologies developed in Wise’s lab to be fundamental to the future of short-pulse fiber laser products, so CCTEC decided to license its patents on the technology nonexclusively. Due to this decision, companies developing lasers for a wide variety of applications will be able to incorporate self-similar or ANDi lasers. Products based on the self-similar laser are already being sold, and ANDi lasers will soon appear in products for security screening, imaging, machining, microscopy, and a host of other applications. The Next Challenge “To get from a solid-state laser to the same performance in fiber, there are some major scientific challenges,” Wise remarked, “and that’s what attracts us. It takes new concepts in pulse formation to make a stable pulse of very high energy that will propagate in fiber.”

Wise and his research group recently developed a fiber laser that generates pairs of synchronized pulses of different colors. These are needed for a new microscopy technique called coherent anti-Stokes Raman scattering (CARS) microscopy. A patent application for the fiber laser has been filed, and Cornell transferred the technology to a German company, which is building a prototype product. With collaborators at Harvard who developed the microscopy technique, Wise’s lab has used the new fiber laser to image mouse brain, cells, and more.

Now that those new concepts have yielded a fiber laser that outperforms solid-state lasers in most areas, the next challenge for Wise and his group is to build integrated fiber systems. They want to answer the question, can we retain a major fraction of the performance we have achieved to make a cheap, integrated, and robust system? I, for one, am betting that they can. Scott S. Macfarlane Former Technology Commercialization Officer CCTEC Director Research Technology and Commercialization SUNY Upstate Medical University


u e




Economic Development

Σ Σ 125

Robert Barker/CU

Through the Lens of Opinions Appinions Inc. How influential are opinions on the web?

Appinions Inc. is a Manhattan-based social media startup that provides software services that let people see the web through the lens of peoples’ opinions. Behind the services is a natural language understanding platform, based on over a decade of research by Claire Cardie, Computer Science, and her students.

ÎŁ 126

Cardie’s research group had been developing statistical machine learning methods to identify opinions and other subjective language in online text, such as news and magazine articles and radio and television broadcasts. Their techniques were designed not only to identify where opinions were mentioned, but also to characterize the opinions according to topic, opinion holder,

and sentiment of the opinion (positive, negative, or neutral). Cardie, cofounder and chief scientist at Appinions (originally named Jodange), is only the first of an unlikely succession of Cornell connections that helped to build the company.


Economic Development

how best to correlate their opinions to outcomes over time.”

In 2006, serial entrepreneur Larry Levy, Appinions’ cofounder and CEO, sold his latest tech company, the Semagix Group, to Warburg Pincus. He was looking for a new business venture in the technology area when he read about Cardie’s sentiment analysis research in the New York Times.

Since then, Levy has been teaching guest lectures each semester in the entrepreneurship classes of David BenDaniel, Johnson Graduate School of Management, one of the faculty behind Big Red Ventures. Johnson School students have engaged in projects to identify emerging market trends in the social media area. One Johnson

Levy contacted Cardie numerous times by phone and by email to discuss the research and determine its commercial viability, but Cardie wasn’t interested in commercialization. Levy persisted, and Cardie agreed to meet him in the Duffield Atrium for lunch one day in December 2006.

In the past year, Appinions expanded its services into the social networking space. It released its Influencer services to find and track those individuals who are most likely to influence a trend, idea, product, or brand. Cornellians were again important in this business development.

“Poor Larry,” smiles Cardie. “He drove from the city out to Ithaca on a very snowy day just for lunch. We had sandwiches from Mattins, and I told him all about fine-grained opinion analysis, the natural language processing techniques behind it, and why it wasn’t yet ready for commercialization.” Levy asked lots of questions and described his thoughts on the many ways that opinion analysis might be used in real-world applications. “Larry asked great questions from the beginning and saw immediately how broadly important the technology could be,” she said. “I couldn’t help but get a bit excited.” Levy and Cardie had more discussions in the following months. When one of Cardie’s former PhD students, David Pierce, decided to leave his principle research scientist position at General Dynamics in Buffalo to find work closer to his family in the New York City area, Levy and Cardie decided that the time was right to launch the company. In late spring of 2007, Pierce came on board as the company’s chief technology officer to lead the technology transfer efforts. In subsequent years, Appinions drew on many Cornell resources. Appinions has hired Cornell computer science undergraduates as summer interns. One of the company’s full-time software engineers is a master of engineering student from Cardie’s classes. Appinions also received funding from the Johnson School’s Big Red Ventures in 2009.

School student spent two months at the company to join its marketing effort. BenDaniel continues to provide advice and guidance to Levy during his visits to Cornell. In the past year, Appinions expanded its services into the social networking space. It released its Influencer services to find and track those individuals who are most likely to influence a trend, idea, product, or brand. Cornellians were again important in this business development. Vlad Barash and Shaomei Wu, two information science PhD students of social networking expert Michael Macy, Sociology/Information Science, consulted in the design of Appinions’ opinion-based social networking algorithms. The Influencer products rely on Appinions’ extensive database of opinions—drawn from blog posts, newsgroups, Twitter, Facebook, company reports, news articles, and radio and television broadcasts—to identify the key influencers creating content on specific topics, as well as the influencers attracting the most attention.

Appinions’ customers include major news organizations, like the Economist, and publishers, like Cengage, which use the Appinions’ opinion and influence platforms to provide readers with relevant and

interesting online content from across the web or specific sources. Customers also include marketing companies like repriseMedia, with clients including American Airlines, Hyundai, Sharp, USPS, and Verizon, and in-house marketing departments of companies like Dell Computers, which use the platform to alert clients when media mentions occur and to determine how influential these opinions may be. Appinions had already raised $3.6 million from private investors. In addition, Appinions announced in July 2012 that the company has secured $3 million in funding to “fuel its sales and marketing efforts.” Kristin Morgan Director of Client Services Appinions Inc.


According to Levy, “The relationship between opinions, opinion holders, and topics is one of the key pillars of all our applications. By identifying peoples’ opinions and sentiments about key topics over time, we can offer an in-depth understanding about who is worth listening to and

Σ 127 Photos in Economic Development: Frank DiMeo unless otherwise noted

Research in Focus

Visualizing the Invisible Frank DiMeo

As I review the many exciting advances presented in this issue of Connecting with Cornell on ways of seeing, I am stuck yet again by the extraordinary breadth and depth of Cornell research. Our investigators are inventing and uncovering new tactics for visualizing and seeing the never-before-seen in so many ways. And certainly to see in new ways is often the first step to discovering the new. As Jon Kleinberg states, “Science always advances when we can take things that were once invisible and make them visible.” Cornell research is not only at the cutting edge that now enables us see the once invisible, but also in consistently sharpening the clarity of our vision of both the physical universe and the richness of human interactions. Cornell researchers invent technologies, such as spectrographic imaging scanning tunneling microscopy, enabling them to see the activity of individual electrons in materials, which can lead to new electronic materials; and multiphoton endoscopy that lets physicians directly see cancerous tissue in the body without the delay of biopsies. They create insightful works of art and scholarship that open our eyes to new ways of seeing humanity—who we are. They use computational sciences— applying them vigorously to diverse disciplines—to answer long-standing questions about human behavior and social interactions, human ancestry and genetic inclinations, and human visual perception. This Cornell research is opening up new fields, such as computational social science to accelerate the study of social structures and human behavior, and silicon photonics, which is enabling researchers to create all-optical information systems, supercomputers on chips, and will lead to new inventions we have not yet imagined. Cornell research is consistently moving us further into discovery in the realms of the currently invisible and unknown. Whether members of our outstanding faculty are updating critical facilities such as the Biotechnology Resource Center with its host of high-end biotechnologies to further research in the life sciences, or starting up companies like Appinions (unveiling the influence of opinions on the web), Picoluz (making photonic devices for things like light meters and amplifiers), and Kphotonics (making ultrafast lasers available for research and educational uses), they are revealing the once invisible. Much of this fantastic research happens because Cornell faculty have no hesitation in reaching across disciplines to bring together areas of research as never before—to invent, create, and discover—generating benefits in unforeseen ways. In this issue of Connecting with Cornell on ways of seeing, we learn how our researchers are bridging physics and biology, medicine and physics, computer science and sociology, engineering and psychology, physics and agriculture, law and computer science, art and science, and more to create advanced ways of seeing. This powerful, boundary-free approach is in our DNA. This trait of collaboration is what enables Cornell researchers to excel, by delivering solutions to life’s toughest challenges and by improving human well-being. This is Cornell research.

Robert A. Buhrman Senior Vice Provost for Research


Non-Profit Org U.S. Postage PAID Cornell University

Connecting with Cornell

2012 VOL. 23, NO. 1 Published by the Office of the Vice Provost for Research 222 Day Hall Cornell University Ithaca, New York 14853–2801 TEL: FAX:


(607) 255-7200 (607) 255-9030 SENIOR VICE PROVOST FOR RESEARCH



Catherine E. Long EDITOR

Ernestina Snead COPYEDITING

Englund Literary Services EDITORIAL ASSISTANT


Kelly S. Strickland DESIGN

Change Service Requested

Cornell University is an equal-opportunity, affirmative-action educator and employer.

222 Day Hall Cornell University Ithaca, New York 14853-2801


Connecting with Cornell (23.1) — Ways of Seeing