BioLAB Business Winter 18/19

Page 1


Narrowing the gap between human and machine learning


Alán Aspuru-Guzik has high hopes for AI







Canadian Publications Mail Product—Agreement 40063567


Setting science in motion to create a better world At AvantorÂŽ, a new era has begun. From breakthrough discovery to agile delivery of advanced products and services across

the globe, Avantor manufactures and distributes high-quality products, services and solutions to professionals in the life sciences and advanced technologies industries.

VWR, part of Avantor provides an integrated, seamless purchasing experience optimized for your success. Learn how the new Avantor is moving forward at Shop now at Š2018 Avantor, Inc. All rights reserved.


company profile




Predicting the disease is now possible years before symptoms appear




Canada’s Research Chair in Machine Learning is working to “narrow the gap” between human and machine logic





Montreal’s AI hub strengthens as MILA moves closer to other industry leaders




Alán Aspuru-Guzik wants to ignite a revolution with AI and robotics – and maybe, save the world

CAMH and Dell EMC unveil their neuroinformatics platform, a superstructure of computation











Arlene Dickinson helps small companies do big things







Improve your cell culture now

Culture of Tomorrow The new CellXpert® C170i CO2 Incubator

Are you looking for a 170 L class CO2 incubator that provides flexibility for the future, makes monitoring and documentation easy, and provides optimized growth conditions, even for your sensitive cells? An incubator that also saves money and is produced to the highest standards of quality?

> Stay flexible and upgrade your device later (e.g., O2 control) > Fast temperature and CO2 recovery in less than 5 minutes without overshoot > Up to 25 % more usable space, easy cleaning, vibration and turbulence protection with fan-less design • 800-263-8715 132.A1.0104.A Eppendorf®, the Eppendorf Brand Design, and CellXpert® are registered trademarks of Eppendorf AG, Germany. All rights reserved including graphics and images. Copyright © 2018 by Eppendorf AG.



Jana Manolakos Keren Stephenson David Suzuki


Gabrielle Cole


BioLab Business is published 6 times per year by Jesmar Communications Inc., 30 East Beaver Creek Rd., Suite 202, Richmond Hill, Ontario L4B 1J2. 905.886.5040 Fax: 905.886.6615 One year subscription: Canada $35.00, US $35.00 and foreign $95. Single copies $9.00. Please add GST/HST where applicable. Bio Business subscription and circulation enquiries: Garth Atkinson, Fax: 905.509.0735 Subscriptions to business address only. On occasion, our list is made available to organizations whose products or services may be of interest to you. If you’d rather not receive information, write to us at the address above or call 905.509.3511 The contents of this publication may not be reproduced either in part or in whole without the written consent of the publisher. GST Registration #R124380270.


BioLab Business is a proud member of BIOTECanada and Life Sciences Ontario


This is my second issue as editor of BioLab Business, and you’ll probably notice some big changes. First, we’ve done away with the “flip” format and combined the two names into one – I’m a huge advocate for streamlining! And with the biotechnology and lab fields working so closely – often overlapping – it makes sense to merge the magazines. Another exciting change is the addition of Canadian Food Business in each issue; we’ll still include a larger 16-page section twice yearly, but our audience shows a hunger for this topic (forgive my pun!), so we didn’t want to make you wait months between issues. The most exciting news of all is our redesigned website: Knowing that our readers are tech-savvy and likely want to read articles on-the-go, we embarked on the mission to create a more user-friendly platform to reach a wider audience. As anyone who has relaunched a website knows, this is not a small job – so you can expect to see added content and features as the year unfolds. Be sure to visit often! With this first issue of 2019, the topic we are focusing on is artificial intelligence. AI is becoming as essential to the business and science worlds as oxygen is to our bodies. While we accumulate unimaginable quantities of data thanks to our connected/computing world, AI is the key to making sense of what might be otherwise hidden quantities. In January, when Toronto’s SickKids hospital announced the first-of-itskind “chair in biomedical informatics and artificial intelligence” – Dr. Anna Goldenberg, a senior scientist in genetics and genome biology – the hospital cited the fact that its intensive-care research has collected two trillion data points since 2013. Considering that none of us can count that high, it’s clearly the task of a computer to analyze and distill the pertinent findings. As we explored the many dimensions of this topic, a constant theme cropped up: Canada is a leader in the field. Not only can we claim rights to the “godfather of AI,” Geoffrey Hinton, but our country continues to forge the AI pathway in many directions, from research and healthcare to self-driving technology and big data analytics. This year, Montreal’s Element AI was the only Canadian company to land on the annual “AI 100,” published by CB Insights; however, in an industry that’s moving at the speed of light, I think there’s a good chance that “the tortoise and the hare” might come into play. The race isn’t over, and Canada has a strong foothold. Do I believe AI is the next frontier in science? Absolutely. Do I believe it can solve as many problems as promised? That remains to be seen – but it can certainly make a lot of our jobs easier, and it is already creating a more insightful vision of the world in which we live. How we use Popi Bowman that information is beyond the “mind” of a MANAGING EDITOR computer, and in our hands.


PUBLISHER & CEO Christopher J. Forbes

Publisher of BioLab Business Magazine Printed in Canada







Science writers have one of the greatest jobs on Earth. They bear witness to the fantastic world of discovery, offering a bird’s-eye view of the fascinating swirl of ideas, knowledge and exploration that helps us understand our beautiful blue planet and its place in the cosmos. During a recent interview with Dr. Alán Aspuru-Guzik, Canada 150 Research Chair in Theoretical and Quantum Chemistry and faculty member of the Vector Institute, I can admit my mortal brain was at times too small to fully grasp some of the notions he was proposing in the fields of artificial intelligence and quantum computing. Dr. Aspuru-Guzik opened up a tiny crack in a cosmic portal that let me peek inside this jaw-dropping and complex world (feature on p.34). It’s a world where computers no longer operate on a simple binary, true or false system, but rather in a quantum dimension where everything was true and false at the same time, a paradox à la Schrödinger. It’s the soil, cultivated by scientific collaboration, from which artificial intelligence and self-guided robots, who communicate to each other, emerge – and it’s here where the future of our planet lies, says Dr. Aspuru-Guzik. You can’t save the world alone, he suggests, referring to researchers in scientific hubs around the world – and even science writers like me – as “superheroes” of the Justice League. That’s why the federal government’s $125-million CIFAR PanCanadian Artificial Intelligence Strategy is so critical to establishing a framework for responsible scientific partnerships, including the nation’s three AI institutes: the Alberta Machine Intelligence Institute (Amii) in Edmonton, Mila in Montreal and the Vector Institute in Toronto. Canada’s ramped-up investments in AI technology are part of the reason Aspuru-Guzik left Harvard last year to work in our nation’s AI community; he sees Canada as a global leader in neural networks and deep learning. When you consider Canada’s GDP ranks tenth behind the U.S. and China, it’s pretty impressive that our nation is pushing forward in the global race for AI dominance. We’re gaining, thanks to the talent we’ve attracted. JF Gagne’s 2018 Global Talent Report identified only 22,000 PhD-educated researchers in the entire world who are capable of working in AI research. The U.S. had the highest concentration of these, with 9,010 researchers, followed by the UK with 1,861. Canada ranked third, with 1,154 researchers. Canada’s success at attracting the best and brightest minds is the reason Google set up a deep learning lab in Montreal, and in the process seduced Dr. Graham Taylor, Canada Research Chair in Machine Learning, expert in algorithms and a rising star in the AI community, to join them for his sabbatical from Guelph University. (More about Dr. Taylor on p.20.) AI offers new insights for the scientific community, from Canada’s publicly funded healthcare systems to energy research – and along with that come enormous datasets, goldmines of information for researchers, like those at the Centre for Addiction and Mental Health (CAMH) in Toronto. One of the teams at CAMH is studying algorithms that can accurately predict if a person will get Alzheimer’s disease (more on p. 17). Like Thor’s hammer, AI is a powerful tool – and as Spiderman will tell you, with great power comes great responsibility.

Only 22,000 PhDeducated researchers in the entire world are capable of working in AI research. 9,010 U.S. 1,861 U.K. 1,154 Canada




Dr. David Suzuki is a scientist, broadcaster, author, and co-founder of the David Suzuki Foundation. Ian Hanington is Senior Editor, David Suzuki Foundation. Learn more at

he digital revolution is breaking new ground every day. Technology has a way of doing that. I remember when Hewlett-Packard introduced its first “laptop” computer, which stored a page and a half of writing. It revolutionized my life as a newspaper columnist, allowing me to write on planes or in a tent and submit articles through a phone. I never imagined the steady advances that would lead to today’s powerful laptops, tablets and handheld computers. Once while filming in a remote B.C. forest, I wanted to pan from the roots of a cedar tree along the trunk to the top in a single shot. After spending hours rigging wires and pulleys and struggling to keep the heavy camera from swaying as it rose, our crew gave up in frustration. Recently, we used a light GoPro camera mounted under a drone to get a spectacular high-definition shot in a few minutes! The first time I opened YouTube, I was looking for a video of the astounding phenomenon of mucous secretion by a hagfish, a primitive marine animal. To my surprise, I found several postings, and as I chose one, a list of several others that might be of interest popped up. Two hours later, I realized I’d been sucked in by an incredible range of films. When I first heard about virtual reality, I was invited to put on the goggles and experience it. Crude as those first images were compared to what’s available now, I was immersed in the scenes. It was impressive and exciting, but I suggested that people should be wary of unintended consequences, because virtual reality could eventually appear better than reality. With virtual reality, people could race a car

and experience all the heart-thumping adrenalin of the real thing, then crash and walk away unharmed. We could have a showdown with a gunslinger, lose and fight again. We could indulge in the kinkiest sex without exposure to sexually transmitted infection or other consequences. Why go for the real experience when a virtual one would be risk-free? During a visit to Montreal, I had the opportunity to watch the latest iteration of the digital revolution: images in 3D, HD and 360-degree wrap-around. It was mind-boggling. I swam with whales and zoomed through a forest, listening to actual sounds, along with music and narration. As I watched a spectacular mountain forest, a train suddenly appeared, splashing across a lake and then coming straight at me. As my body responded to the all-too-realistic locomotive, it reached me and exploded into a thousand birds that took off in a glorious cloud. Computer graphics melded seamlessly with actual footage that generated scenes far exceeding reality. I’ve been intrigued by the possibility that this technology could enable people to have such incredible experiences with whales, fish and other animals that we would no longer feel the need to imprison animals in aquaria and zoos. People wouldn’t even need to journey to exotic places to see wildlife in their habitats. I have no doubt virtual reality is going to have a huge impact. We’re just beginning to recognize its potential. But as with all new technology, there will be unintended repercussions, the greatest of which will be further estrangement from nature. Studies show that because people evolved out of nature, we need that connection with the natural world for mental and physical well-being. Author Richard Louv categorizes a suite of childhood problems — including bullying, attention deficit disorder and hyperactivity — as “nature deficit disorder”, induced or worsened by too little physical exposure to nature. The average Canadian kid today spends more than six hours a day glued to a screen — mobile phones, computers, televisions — and less than eight minutes a day outside! That’s one reason why the David Suzuki Foundation encourages people to get outside for 30 minutes a day with its One Nature Challenge. Some proponents claim virtual reality will stimulate children to spend more time outside. But why bother when the virtual world seems better than the real one? I’m sure innovation and creativity will continue to drive the technology to new frontiers. I’m just as sure there will be enormous unexpected and damaging consequences if we aren’t careful.





ROBOTICS AIDE IN DEVELOPING NEW MATERIALS TO IMPROVE ENERGY EFFICIENCY SCIENTISTS GROW PERFECT HUMAN BLOOD VESSELS IN A PETRI DISH Josef Penninger – the Canada 150 Research Chair in Functional Genetics, director of the Life Sciences Institute at UBC and founding director of the Institute for Molecular Biotechnology of the Austrian Academy of Sciences (IMBA) – is senior author for a study published in Nature that resulted in a ground-breaking model: threedimensional human blood vessel organoids grown in a petri dish. When the organoids were transplanted into mice, they developed into “perfectly functional” human blood vessels, including arteries and capillaries. The discovery illustrates that it is possible to not only engineer blood vessel organoids from human stem cells in a dish, but also to grow a functional human vascular system in another species. Its application has possibilities in the treatment of vascular diseases, such as diabetes.

In late 2018, Canada’s Minister of Natural Resources, the Honourable Amarjeet Sohi, announced an $8-million investment to develop a first-of-its-kind robotic platform, called Ada – the Autonomous Discovery Accelerator (, also named after mathematician/writer Ada Lovelace – which uses artificial intelligence to allow Materials Acceleration Platforms (MAPs) to get smarter and faster after every experiment, as they learn and optimize processes using custom-made machine learning algorithms. Developed by the University of British Columbia as part of an international collaboration, Ada was highlighted by the World Economic Forum as an example of “how AI is revolutionizing the world of clean energy materials.” The autonomous lab tests materials at high computing powers, aiming to make solar panels more resilient and turn carbon dioxide into useful fuels.


investment to develop a first-of-its-kind

robotic platform,

called Ada



Left: Project Ada robot laboratory



DEEP GENOMICS TACKLES NEURODEGENERATIVE DISORDERS VIA AI In Toronto, Deep Genomics has built a biologically accurate machinelearning platform that supports geneticists, molecular biologists, chemists, toxicologists and drug developers in the identification and development of genetic medicines. The company is hard at work after announcing a $10 million investment last spring to expand its preclinical platform and to develop therapies for metabolic and neurodegenerative disorders. Its three-year plan aims to unlock new classes of anti-sense oligonucleotide therapies that were previously inaccessible, and advance them for clinical evaluation. “We have allocated 50 per cent of our drug discovery pipeline to finding new therapies for genetically defined metabolic and neurodegenerative disorders,” said Dr. Frey, CEO and scientific founder of Deep Genomics. In a collaboration with Wave Life Sciences (based in Cambridge, Mass.), they are focusing on the treatment of genetic neuromuscular disorders.

“We have allocated 50 per cent of our drug discovery pipeline to finding new therapies for genetically defined metabolic and neurodegenerative disorders.” Dr. Frey, CEO and scientific founder, Deep Genomics

Gelsolin destroys filaments within the heart's cells. PI3K alpha is able to halt this process.

PRICK-FREE GLUCOSE MONITOR DEVELOPED A research team at the University of Waterloo has combined radar and AI technologies to detect changes in blood glucose levels without the need for invasive means. “We want to sense blood inside the body without actually having to sample any fluid,” explained Waterloo Engineering professor (and team leader) George Shaker. “Our hope is this can be realized as a smartwatch to continuously monitor glucose.” The research involved collaboration with Google and German hardware company Infineon, which jointly developed a small radar device and sought input from select teams around the world on potential applications. The system at Waterloo uses the radar device to send high-frequency radio waves into liquids containing various levels of glucose, and receive radio waves that are reflected back to it. Information on the reflected waves is then converted into digital data for analysis by machine-learning AI algorithms developed by the researchers. Initial tests achieved results that were 85 per cent as accurate as traditional blood analysis.


Approximately 20 per cent of heart failure cases are the result of dilated cardiomyopathy. An international research team led by Gavin Oudit – a professor of cardiology at the University of Alberta and director of the Heart Function Clinic at the Mazankowski Alberta Heart Institute – discovered a key molecule that binds to and suppresses gelsolin, an enzyme that is linked to the destruction of heart tissue in these cases. Oudit explains that biomechanical stress activates gelsolin, which then destroys the filaments within the heart’s cells. The molecule, named PI3K alpha, is able to halt this process.



UWATERLOO CHEMISTS USE AI TO INTERPRET DMS RESULTS Artificial intelligence is helping discover new drugs faster, more affordably and more efficiently. In a study published by Nature Communications, chemists at the University of Waterloo used AI to interpret results acquired by differential mobility spectrometry (DMS) to predict drug properties. DMS is a technique that analyzes molecules based on their response to an electrical field and condensation-evaporation cycles. “AI has reduced the analysis time and made the process general and more efficient,” said Scott Hopkins, professor of chemistry at Waterloo. “With the introduction of machine learning, we can examine numerous types of drugs simultaneously. This doesn’t stop at drug molecules; we can pretty much study any molecular system this way.”

“AI has reduced the analysis time and made the process general and more efficient”

VECTOR INSTITUTE OFFERS SCHOLARSHIPS FOR AI MASTER’S PROGRAM Up to 115 scholarships will be awarded to students who pursue a full-time AI-related master’s degree in the 2019–20 academic year. Scholarships (valued at $17,500 for one full year) will be awarded to students in both core technical programs and complementary fields such as business and healthcare. Interested candidates can find eligibility and application details online at: Nominations must be received by April 5.

Scott Hopkins




On April 10–11, the first-ever World Summit AI Americas visits the St. James Theatre in Montreal. The two-day event will be packed with over 1,000 attendees and 60 speakers, including the minds behind companies such as Pinterest, Google, Amazon, LinkedIn and more. Many activities are planned, including panels, coding classes and tech workshops. For more information, visit:


GOVERNMENT OF CANADA INVESTS HEAVILY IN AI DEVELOPMENT Several initiatives were recently announced as the Government of Canada reaffirms the importance of AI as an engine for economic growth. In December, the Honourable Navdeep Bains, Minister of Innovation, Science and Economic Development, announced an investment of nearly $6.3 million into six Al-based businesses, all located in Montreal. “Artificial intelligence is transforming all industries and sectors, opening up more opportunities for Canadians,” Bains said. “It’s exciting to see the vitality and creativity of these six companies that, each in its own way, are helping to strengthen Canada’s role as a global leader in innovation and creating highly specialized jobs that play an essential role in the growth of Canada’s digital economy.” The business receiving the most money (a $5 million loan), Element AI attracted worldwide attention in 2017 when it secured more than $100 million from a group of U.S.-based investors. Co-founded by AI scientist Yoshua Bengio, the company is launching its first software products this year. Element AI plans to create around 900 new jobs as it expands, while the other companies receiving support will generate almost 80 jobs. Nearly $1.3 million is being distributed among five other AI-related Montreal businesses: Imagia, Keatext, ARA Robotic, C2RO Cloud Robotics and Roof Ai. This announcement followed a significant commitment of $25 million for the Creative Destruction Lab (CDL), a non-profit organization founded in 2012 and based at the University of Toronto. CDL merges science-based projects with business expertise to help young companies thrive. With this investment CDL plans to create 125 jobs, while supporting more than 1,300 ventures across Canada

over four years; it’s estimated these projects could create up to 22,000 new jobs. Along with boosting engagement of young women in STEM, the CDL projects will focus on supporting business ventures that harness emerging technologies such as AI, clean tech, energy, health, smart cities and quantum technologies. As part of the $1.26-billion Strategic Innovation Fund, this investment is one of many; earlier this year, a Health and Biosciences funding competition was narrowed down to a shortlist of nine companies (including Imagia, XCO Tech, Precinomics Health Solutions and the Sunnybrook Research Institute), with winners to be announced in May. The nonrepayable funding won’t exceed 50 per cent of eligible costs, except for academic institutions and networks, which are eligible for 100 per cent funding. Almost all of the finalists incorporated AI, or some form of data management, into their project plans. Meanwhile, the AI-Powered Supply Chains Supercluster (SCALE.AI) – part of the Government of Canada’s $950-million Innovation Superclusters Initiative (matched dollar-for-dollar by the private sector) – will bring multiple industries together to build intelligent supply chains using artificial intelligence and robotics. Based in Montreal and spanning the Quebec-Windsor corridor, SCALE.AI is expected to create more than 16,000 jobs and generate more than $16.5 billion in GDP impact. For more information about the participating companies and projects, visit:

“Artificial intelligence is transforming all industries and sectors, opening up more opportunities for Canadians." Honourable Navdeep Bains, Minister of Innovation, Science and Economic Development


Investment of nearly $6.3 million has been made into six Al-based businesses, all located in Montreal.




Researchers aim to characterize the kinematics and past rupture history of several important fault systems in northern Baja California. Photo credit: SOI/Monika Naranjo Gonzalez



Artificial intelligence is powering many new discoveries, from the ocean floor to outer space. Late last year, a newly discovered deep-sea hydrothermal vent field was digitally mapped and sampled during a multidisciplinary Schmidt Ocean Institute research expedition, which used Monterey Bay Aquarium Research Institution’s (MBARI) Dorado autonomous underwater vehicle and a remote-operated unit equipped with multibeam sonar, scanning laser Lidar and stereo photography. Detailed maps and samples enable the study of microbial extremophiles involved in methane and hydrocarbon metabolism, which are unique to this site, where multiple hydrothermal calcite mounds up to 25 meters high are venting fluids at temperatures as high as 287°C. Many previously unknown species were identified. “The deep ocean is still one of the least explored frontiers in the solar system,” said Principal Investigator Robert Zierenberg. “Maps of our planet are not as detailed as those of Mercury, Venus, Mars or the moon, because it is hard to map underwater. This is the frontier.” For more about the mission, visit: Another AI-driven ocean study (published in Marine Micropaleontology, vol. 147) developed a highly accurate system for identifying microscopic marine organisms: specifically, six species of foraminifera, or forams, which have existed for more than 100 million years. A specialized microscope creates a detailed photo model, which is then analyzed and identified by the AI system. “Expert” human analysis averaged 63 per cent accuracy, while the computer achieved better than 80 per cent. Plans are underway to train the AI system to recognize at least 35 species. The research team received a grant from the National Science Foundation to build a fully functional AI-equipped robotic system, starting this January. In other groundbreaking research, Jason Nordhaus – an astrophysicist at the Rochester Institute of Technology (RIT) who is partially funded by a three-year grant from the NASA/ Space Telescope Science Institute – developed a system of complex, 3D supercomputer

Maps of our planet are not as detailed as those of Mercury, Venus, Mars or the moon, because it is hard to map underwater. This is the frontier. Robert Zierenberg

Top: The Autonomous Underwater Vehicle is equipped with four mapping sonars that operate simultaneously: a swath multibeam sonar, two sidescan sonars and a sub-bottom profiler. Photo credit: SOI/Monika Naranjo Gonzalez Centre: Dr. Dawn Cardace processing samples in Falkor's wet lab. Photo credit: Courtesy of Schmidt Ocean Institute.

algorithms that have pinpointed the existence of previously undiscovered planets and celestial bodies associated with dying stars. “This helps us understand the fate of our own solar system, the fates of other star systems in the galaxy, and improve our understanding of how stars and planets interact,” Nordhaus explains. He is also a member of RIT’s Center for Computational Relativity and Gravitation, whose simulations helped confirm the breakthrough detection of gravitational waves from binary black holes in space. But even “new” discoveries can be found in old data. The SETI Institute in Berkeley, California, has been using IBM Cloud and AI algorithms to analyze over 20 million signals captured by radio telescopes, using machine learning to identify and flag significant events. An AI system developed by its researchers recently identified more than 70 “fast radio bursts” (FRBs) originating from a galaxy approximately three billion light years from Earth, by searching through 400 terabytes of data that had been previously collected. While most FRBs occur only once, this is the first case where multiple pulses have been generated by the same location. What causes these frequencies is not determined, but SETI continues to collect and share data from space in hopes that these mysteries will someday be solved. “These results hint that there could be vast numbers of additional signals that our current algorithms are missing, and clearly demonstrate the power of applying modern data analytics and AI tools to astronomical research,” says SETI Institute President and CEO Bill Diamond. “Applying these techniques in the search for evidence of extraterrestrial technologies, or technosignatures, is incredibly compelling, together with addressing the tantalizing phenomena of FRBs.” Meanwhile, newly formed Hypergiant Galactic Systems (a subsidiary of Hypergiant Industries) promises to “expand humanity into the stars with machine intelligence,” focusing on AI-driven aerospace and astronautic software and hardware products, with services including orbital system consulting and deployment. “Our technology is leveraging multiple frequencies of non-visible spectrums to produce images based upon the reflectance of energy from the surface, revealing information hidden below visible spectrum,” the company explains. “Coupled with our custom image recognition systems and contemporary information architectures and data stores, we have the ability to classify information as it comes in.” Whether analyzing satellite data or enabling space exploration, this company appears to be on the forefront of technology. Only time will tell what its contributions will be.

- Popi Bowman





PROTEIN ENGINEERING EXTENDS THE LANGUAGE OF IMMUNE CELLS An interdisciplinary research team at the Technical University of Munich (TUM) designed a new immune-signalling molecule that provides the basis for potential approaches in sepsis therapy. Millions of people die each year from sepsis, commonly called “blood poisoning,” which is an overreaction of the immune system; in Germany alone, more people die of sepsis than of AIDS, colon cancer and breast cancer combined. “Using computer models and cell biological experiments, we discovered that a single structurally important amino acid defines whether interleukin27-alpha is released by cells of the immune system,” explains Stephanie Müller, the first author of the TUM study. Interleukins are, essentially, messengers that mediate communication between the cells of the immune system; this particular molecule, interleukin-27-alpha, is only produced in mice. “That gave us an idea about how we can engineer novel human interleukin proteins that are released by cells, so that we can produce them biotechnologically,” Müller says. The team prepared the modified interleukin in the laboratory, and tested its biological functions, with very encouraging results: The engineered messenger molecule is recognized by human cells. Initial analyses indicate this protein could balance an overreaction of the immune system, making it a promising candidate for sepsis therapy. As of press time, a patent for the new protein is pending.



Using computer models and cell biological experiments, we discovered that a single structurally important amino acid defines whether interleukin-27-alpha is released by cells of the immune system. Stephanie Müller, Technical University of Munich (TUM) research team


In a list of “Five young companies making an impact on the world,” included Cognetivity Neurosciences. The London-based company (with an office in Vancouver) is aiming to improve dementia diagnosis rates by administering visual tests to challenge the brain, using AI to analyze whether or not that person is at risk of developing dementia later in life. This early and incredibly accurate diagnostic process can save money (around $118,000 per patient), as well as increase the possibility of effective treatment to better help those living with the disease. Dementia affects nearly 10 million people a year, meaning someone develops dementia roughly every 3.2 seconds; the World Health Organization reports that the number of people living with dementia worldwide is estimated to reach 75 million by 2030. Alzheimer’s is one of the most common forms of the disease, contributing to 60-70 per cent of cases. We explore other AI-driven diagnostic systems for Alzheimer’s on p.17.

Dementia affects nearly 10 million people a year

WEF URGES CAUTION DURING THE “ROBOT REVOLUTION” The World Economic Forum (WEF) identifies AI as “the software engine that drives the Fourth Industrial Revolution.” In 2017, the WEF established its first Center for the Fourth Industrial Revolution in San Francisco, to act as a hub for discussing how science and technology policies “can benefit all in society.” A large part of that conversation centres around how to steer the development and implementation of AI: “The Fourth Industrial Revolution is reshaping industries, challenging existing regulatory frameworks and redefining what it means to be a human,” says Murat Sönmez, Member of the World Economic Forum Managing Board and Head of the Center. “We need to urgently develop policy norms and frameworks and apply these innovations to ensure their benefits affect us all.” The Fourth Industrial Revolution was also the topic for this January’s WEF annual meeting in Davos, Switzerland. Recent concerns include gender inequity within the field; less than a quarter of roles in the industry are filled by women, often resulting in a systemic bias – for example, resulting in the choice of a woman’s voice as a “service” role (ie. Alexa, Siri). On a lighter note, the WEF estimates that AI will create more jobs than it destroys, a surplus of almost 60 million worldwide by 2022.

The many faces of the robot revolution

Adoption among companies by 2022

Humanoid Robots

Stationary Robots

Aerial and Underwater Robots

Non-humanoid Land Robots





(52%) Oil and Gas

(42%) Automotive, Aerospace, Supply Chain

(35%) (53%) Financial Automotive, Services Aerospace, and Investors Supply Chain

First movers

Source: Future of Jobs Report 2018, World Economic Forum

Rate of automation Division of labour as share of hours spent (%) Human



2018 2022


58 48

Source: Future of Jobs Report 2018, World Economic Forum

29 42 52
















Canadians are living with AD or another form of dementia


he once vibrant and detailed brushstrokes of American artist William Utermohlen began to disintegrate with the onset of Alzheimer’s disease (AD), a diagnosis he received in 1995. By 2000, he was in a nursing home, unable to care for himself. Utermohlen might have staved off deterioration, and continued to paint for several more years, if a new AI-powered algorithm capable of early detection had been available at the time. Just such an algorithm was the subject of a study published last September in PLOS Computational Biology. Authors Mallar Chakravarty, a computational neuroscientist at McGill University in Montreal, with colleagues Nikhil Bhagwat and Joseph Viviano, from the University of Toronto, and Aristotle Voineskos, from the Centre for Addiction and Mental Health, found an algorithm that can accurately predict Alzheimer’s onset five to seven years before symptoms appear. “At the moment, there are limited ways to treat Alzheimer’s, and the best evidence we have is for prevention. Our AI methodology could have significant implications as a ‘doctor’s assistant’ that would help stream people on the right pathway for treatment,” says Chakravarty. “For example, one could even initiate lifestyle changes that may delay the beginning stages of Alzheimer’s, or even prevent it altogether.” Alzheimer’s disease is believed to be caused by a buildup of sticky amyloid plaques and tau proteins that disable neurons from communicating with each other. One of the big challenges for clinicians is that the disease is mainly detected once it is advanced. The researchers hope their new algorithm could change this by providing an early warning system. According to Canada’s Alzheimer’s Association, over 747,000 Canadians are living with AD or another form of dementia. There is presently no cure, nor can Alzheimer’s be reversed; however, current treatment options and lifestyle choices – for example, eliminating a risk factor such as high blood pressure – can significantly slow the progression of the disease. For their research, the team used more than 1,000 participants, divided according to their state of wellbeing, from normal healthy seniors to those experiencing mild cognitive impairment, and Alzheimer’s disease patients. Data was collected on clinical exams of mental status (MMSE) and AD assessment scales (ADAS-13) over six years, as well as brain magnetic resonance imaging (MRI), and genetic and clinical reports. The data was then


ANOTHER AI APPROACH TO ALZHEIMER’S DIAGNOSIS In Montreal, Optina Diagnostics is exploring the possibility of diagnosing Alzheimer’s by using a hyperspectral camera to analyze a patient’s retina, localizing and quantifying specific biomolecules in the fundus. A characteristic of Alzheimer’s disease is the accumulation of beta-amyloid plaques in the brain; since the retina is an extension of the brain, it provides critical access for identifying beta-amyloid plaques, known as a key biomarker of Alzheimer’s. First, patients undergo a full eye exam by an ophthalmologist and receive an imaging test via Optina Diagnostics’ hyperspectral research MHRC camera; this takes less than a second, during which patients will see a rainbow of colours. For their second visit, patients undergo a PET scan using a beta-amyloid tracer; this scan uses a special dye containing radioactive tracers, which are absorbed by the brain and tissues for diagnostic purposes. Results are sent to a cloud server to be analyzed by Optina’s proprietary software, which identifies whether amyloid plaque deposits are present. In collaboration with Imagia Cybernetics, a healthcare AI company also based in Montreal, the retinal imaging is analyzed via artificial intelligence. “By combining our data-rich images from a large number of very narrowband wavelengths with the deep learning pipelines developed by Imagia Cybernetics, we believe we can accelerate the delivery of novel biomarkers and predict patient outcomes,” says David Lapointe, CEO at Optina Diagnostics. The goal is to revolutionize the diagnostic process, and already their results are promising.


run through the algorithm, which the research paper refers to as a “longitudinal Siamese neural-network (LSN) with novel architectural modules for combining multimodal data from two time points.” The computer uses a complex decision-making process to predict outcomes based on several pieces of information from different points in time and is considered highly accurate. Chakravarty says it’s a bit like the algorithms used by Facebook or Google for identifying people in online images. “We can repurpose those kinds of algorithms to ask whether this person looks like they have Alzheimer’s disease,” he explains. “Do the patterns in their brain, genetics or cognitive profile predict a high likelihood of the disease?” The team is also looking at the algorithm as a means to recruit people for future studies, who might be at risk, and to follow them down the road. They’d like to see more participants in order to fine-tune the algorithms and, as Chakravarty says, “increase our specificity.” “We are currently working on testing the accuracy of predictions using new data,” he explains. “It will help us to refine predictions and determine if we can predict even farther into the future.” The researchers hope the model will help doctors make accurate predictions on cognitive decline not just for Alzheimer’s, but for other neurodegenerative diseases as well.












QA +


f predictive algorithms had been around when Graham Taylor was born, they would have forecast a skyrocketing career, with a trajectory that is still gaining momentum today. Dr. Taylor recently became the academic director of the newly established Centre for Advancing Responsible and Ethical Artificial Intelligence (CARE-AI) at the University of Guelph; he actively promotes entrepreneurship as academic director of an artificial intelligence (AI) incubator called Next AI; and, he’s a member of Toronto’s Vector Institute for Artificial Intelligence. Last September, Taylor was awarded the Canada Research Chair in Machine Learning, and this past November he was recognized by Caldwell and MNP as one of Canada’s “Top 40 under 40.” He’s someone to watch, so we wanted to learn more.



Why is recognition as a “Top 40 under 40” important? It really recognizes the quality of my lab’s work, and that involves nearly 20 people. It is very meaningful for all the students and post-docs in my group, as well as for the University of Guelph, where I have been for six years. The award reflects my interactions with all of these links, including a number of organizations that I collaborate with, like NextAI and the Vector Institute. It’s really a team effort.



As an undergrad and Master’s student in Engineering, specializing in systems design at the University of Waterloo, what started you down the path of machine learning? My parents bought me a computer when I was seven. They were not computer people, and basically just handed me the box. I figured out how to use it and spent so much time with that little, low-power computer. That curiosity continued throughout school. By the end of high school, I didn’t want to pigeonhole myself by taking computer science or computer engineering; I wanted to be more rounded, so I pursued the study of systems. It’s a very broad discipline. There are so many systems in the world – biological, environmental, and even a human is a system. But I also couldn’t get computing out of my “system,” so I ended up migrating towards AI. I was inspired by a sessional instructor, Andrew Brown, I had at the University of Waterloo, who taught an introduction to artificial intelligence. He set up a competition for students

to build intelligent agents to play a board game against each other. Our team ended up winning. It was so much fun that I decided to pursue AI in graduate school. Who is your most admired scientific mind? Someone I actually worked with and really admire is my PhD advisor, Geoffrey Hinton [commonly known as the Canadian “godfather” of AI].

Some people have said that AI is a generalpurpose technology because it can impact so many different fields, but for the same reason, it also becomes challenging.


while I don’t think that in our lifetime we’ll be making machines that are indistinguishable from humans, we can make progress in that direction. I think we can certainly make learning algorithms that are a lot better than the ones now. Can you tell me about your collaboration on camera traps and bio-informatics? This is the work of PhD student Stefan Schneider at the University of Guelph. He studied ecology in undergrad and is now studying computing. What he is looking at is basically automating camera traps to be able to recognize both the species seen in the trap and also individual animals. He is working on data sets like those produced by Snapshot Serengeti, which collects information from pictures taken by camera traps in Africa. This is beneficial to not only understanding ecological patterns, but also fighting poaching. He is also collaborating with other researchers on underwater cameras, trying to understand the social behaviour of octopuses. But they don’t have enough time to watch and annotate these volumes of video, so we are hoping that an algorithm can be more efficient than a human for processing the amount of data coming from the cameras.

What are your current priorities? I’m most excited about making deep learning algorithms that narrow the gap between the way humans learn and the way machines learn. If we can close this gap and improve the algorithm, then we can touch many other applications. And

What about machine learning and health care, specifically your collaboration on patient flow-through in smart hospitals? That is the work of Petros Spachos, a researcher at the University of Guelph working on wireless sensor networks, and a Master’s student, Maeve Kennedy. Their work is looking at the use of beacons in hospitals for patient tracking, which could be helpful in long-term care settings with wandering dementia patients. It’s all about improving the security, privacy and access to care in those facilities. You have been working with time series. What is that? It’s data that has a temporal or sequential element to it, like observing weather over a number of days or the fluctuations of the stock markets. We are looking at building machine learning systems that we can train with time series data. For example, a single screenshot in a video might have millions of pixels. Over time, we would have to store those millions of data points for every split second of the video. It gets very large. Times series means more data, more storage and more-sophisticated algorithms to be able to process all of that. AI can help us in determining what is important to store and what isn’t. It can also help us compress larger amounts of data. You started your own incubator, NextAI. What is happening there? I’m mentoring a number of startups in that program. Last year


The scope of AI technology is very broad, from biology to finance. Where do you even start in deciding on a specific aspect of AI? Some people have said that AI is a general-purpose technology because it can impact so many different fields, but for the same reason, it also becomes challenging. For me, the last six years have been about learning how to choose the things I’m most passionate about, while exploring the space of opportunities in machine learning. In algorithms, we often look for a learning agent to be able to both “explore and exploit,” and what this means is that “exploiting” is doing what’s worked well for you in the past; the quintessential example of this is called “Bandit” theory, which comes from the one-armed bandit in a casino, where if you pull an arm, you can get a payoff. You can keep pulling on that arm and “exploit” it, or you can head somewhere else in the casino to explore the other machines for a potentially even bigger payout. Relating that back to a career, you will realize some successes in your career pursuing a particular application, but there are also other things to explore because you don’t know if you have hit your peak yet. You need to balance these two.



I’m working on a paradigm called reinforcement learning, which is based on the psychology of reward and punishment in helping the machine gain knowledge. Like a robot trying to solve a maze, moving around until it finds its goal – it’s very similar to how we might learn as humans, the trial and error method.

we had teams coming from such countries as the Ukraine, England and India. We are encouraging these people to start a company here and are working with the Canadian government for a streamlined immigration process. You are currently on sabbatical at Google Brain. What are you working on there? There are different paradigms to machine learning. For example, there’s supervised, unsupervised or reinforcement learning. The way we typically build systems is through supervised learning, where there’s a teacher that guides the machine in making predictions. There are other ways of learning that use much less supervision. I’m working on a paradigm called reinforcement learning, which is based on the psychology of reward and punishment in helping the machine gain knowledge. Like a robot trying to solve a maze, moving around until it finds its goal – it’s very similar to how we might learn as humans, the trial and error method.



What does the future hold for AI and its relationship to mankind? The University of Guelph recently established the Centre for the Advancement of Responsible and Ethical AI. The centre aligns campus work around the responsible deployment of AI and is set up along three pillars that aim to drive improvement, to ensure responsible deployment – so they are fair, without bias, transparent and safe, especially if they are physical, like robots, and we want the creators of this work, like the engineers and scientists, to also be accountable. The third pillar is on building applications that do good in the world, and solve big problems, like feeding the world sustainably or protecting the environment. Is there a broader Canadian perspective? The Canadian Institute for Advanced Research is leading the way on that. They have an AI and society program and have collected a number of researchers together in a national consortium to address issues like ethics, policy and safety. The Montreal Institute for Learning Algorithms [MILA, featured on p.26] has its AI4Good program, and there are others. People are really rallying around that concept to make sure AI is used to benefit people, rather than harm them.

Stefan Schneider, a PhD candidate at the University of Guelph, works on a research team focused on the potential for AI systems to re-identify animal individuals from camera traps in order to automate population estimates. Schneider says that "this approach benefits conservation groups worldwide by eliminating the need for tagging animals, which is expensive, labour intensive, and invasive to animal individuals. The ability to promptly identify when ecological systems are in peril can have enormous environmental and economic impacts on the communities which depend on them. AI will also help us classify animal behaviour in their natural habitat, as well as provide insight into seasonal dynamics related to travel direction and when offspring occur most frequently."

Petros Spachos, Assistant Professor, School of Engineering, University of Guelph teamed up with Master’s student Maeve Kennedy and the university’s Canada Research Chair in Machine Learning, Dr. Graham Taylor, to explore the use of inexpensive Bluetooth Low Energy (BLE) beacons for capturing patients' vital signs and locations in smart hospitals. Spachos says that “after capturing the signal, some processing is necessary to extract useful information, which is then stored and used to create user profiles. So, it is a combination of wireless communication with machine learning.” He says that beyond tracking wandering patients with Alzheimer’s disease or dementia, the technology offers many more applications that range from asset tracking to automated check-in.


NEW FOR 2019

Together, We Reach Farther into the Canadian Science Community For over 30 years, BioLAB Business has informed laboratory and biotechnology professionals, while our sister publication Canadian Food Business enhances your reach with Canada’s food and beverage science industry decision-makers.

user-friendly 1 Fresh, website


JAN sletter UA No. 1 RY 0 201 1 9

BioL AB B usin ess Web site CFB Web site Ema il for more in Invit fo rece e a colle ive th a is ne gue to wsle tter

issues of 2 6combined editorial

SUB SCR BioL IBE AB B usin ess Mag CFB azin Mag e azin e

3 64 eNewsletters: BioLAB Business



New s: BIO

Pro duc ts:


NTS JUN Bio E 3-5 m Sum anufa San mit 2 cturin 0 g ww Diego 14 w.b , CA iom ansu JUN Wit mm h it.c DIA E 15-1 alon 2.3 m om Mee 2014 9 HIV e, inte illion n San ting 50th A HIV vacci rnation ew ca nnu ses /AID ne is ally http Diego al , :/ C S u imp reno of HIV A .org /ww wn nder erat In M w /en con ive to ed vac every ay -US .diaho Mel , trol. com cine year .asp me ind Espar JUN n x plem exp glob a za ad G w ia 2014 E 23-2 ent ert Jo ally, "Do n mee ates F ill re oth in ti o u re Con BIO 6 W ti er p se Esp cludin nd n add e fr In ar rev ress Need g on H ation. om h San ventio ternat enti za sa g 50,0 A is ro IV H a at n ll ve in ys th 00 N e ion ia http Diego al dec nce C the Can ew Par vaccin will d le as terv e nee in th ades el e e o enti bio ://con , CA ad ad ord iv Senio re effi .org ve ons d for an U.S. rA cacy in vir inatin ian HIVigm fo search er his /20 nton dvis and fi H r al g an rs H tr IV 14/ . brin vac ials, an disease Office Vaccin IV Vac d dev t post or, Vac g cine -r el e (A ci s discod stron and v CO) Initiati cine D opmen etirem nes at Esp ver g inte accin Annu ve (C evelo t. His ent sp the B cine arza an y. al M HV pmen pre eech ill rnat e re & se EXT se I) ion Ann devel d 12 o FOR al p arch. eeting Resea t?" w ntati at a C o EN on il . aartn reg ual CH pmen ther C H DE OUR an u DC t ersh e says Esparza rch an l be th entitl mu latory VI R& and th adian e ALE d ed ips ltid ND wil more in has w Devel keyno isci affairs, D Mee e inno and in l be AR te o o plin v the novativ rked fo pment Clic ary pharm ting. T ative ternatio key step dialo a an he n s to e rese r Bio k here fo gue d in meeti s bei al exp succ arch r Bus that ternat ng w ng ta erts essf , wil k Esp io il ul ine arza nal o l prov en to o l add ss ress ver ide has rgan Clic adv izatio persp come barri New oca er LAB k here fo ted ns - p ectives them at s to v in ef regula re for ac Bus r thro cisely from the A fect fect. T tions re C g ugh th ine purp and w he Mar overn out e kin search O ss d , his ers oses b ill fund ihuan ing the care of a (LP y er. s) re movinamenta for M supply ed lly gula an g "Th ted to a n chan ical Pu d distr es g by purp e reg Hea ew com e the rposes ibutio w n lth edu oses. ulation Can merci ay Can Regula of med W ca issu tional e are s will ada. al sy ti ch stem adians ons (M ical ca Inte es rela instit comm ange MP nnab led acce te u it rim ss R is by -Ch d to th tions, ted to the way app marih ) cam are n h air, w e rov e o The CM acce ealth ca orkin Canad ed li uana into ef w g CIA ss an ia thro CMC cen for m d re re prov positiv ns ac . IA sed ed ce and gula (MM ughou el id pro ical t tion ers an y with ss can ducnatio AR) the tr its fou n d o ab o fm la an u to n edic w en r patie is for inte nal st the M sfer ding from mem med grity and al ca forc n M ts ber nnab emen , gov ical of th ards an PR sy the t ag ernm pre s hav st is em d em ," v eb io say encies ent, edic best sM al ca pract we hav us Med een w arc on ork e to nnab ices ic Way is in by su day. T al Mar ing al ne, on iju dust h p ry. portin e CMC ana A gside g g th cc e dev IA's m ess R overnm elop issio egula ent tio men n is t, g to pro ns row th an mote d


Brin Vac ging th cine e HIV , Ex per Pande t Sa ys mic to

and 2 Canadian Food Business

HEA D this LIN mon ES th

me MEDIC ans A for L SE port LFIE able S: A U dia gno BC te stic am NE su o X sin f che INT T-GE g ce mis llph ts d by ENSIF NERA one isco Mill IED TIO s. ver ipo N reS DRUG PRO a igm C a a PROD ESS T im U to h CTIO ECHN O elp N bio : Thre LOGIE ma nufa e new S FOR p ctu rers roduc nav ts iga te

Zer oW

ill R equ ire a

Lau n Me ch of M dica l Ca MPR L nna bis eads to Lan dsc Chang ape es

25,504* Circulation

to C ana da's

2.3** 7,974 66,633 Readers per copy $ 37 .5 M

Digital Circulation


Trea il tme lion in Ren nts ewe d Fu ndin g fo r Be tter


17,359*** Circulation

BioLabCFB_19_Rate card.indd 1

of O Ontari of th ntario o Brian ren inking to enco Instit ew men ed p aroun urage, ute has ro work t Diso gram d som rewar receiv e rd ery Integ ers In s inclu of the d, and ed a fi v ca Pro d ra gra ted D tegrate e PON most p talyze e-year m). isco d D com D N ervas s st The is ro ver y P cover etwork ive bra nger mitmen utes OBI w rog y ram Prog (The in dis collab t from tota 2/3, an orks w ora ord Pro ), an ram th l it d incr progra its h h inst d E ), CP vince ers o tion an e Pro fo pLin v ease -NE m co ospit ituti d o k (T T (T f Onta ur tim new ince d re o way sou sts. T al and ns to e. he E he ri Bra rces hese univ fun pilep Cereb o Neu The th s in dp er to en par disco diso la sy In ral rodev ree rd Pal han tner cosity par tform te div very ers af sy N elopgra ce th s tn er n an ted ete sc tributi ers co for dis patie se gro d co fect m D m co o n is il u pe o ons tr nts cov v and ps of exmercial lions o f th stren ibute th ery w f e re thei h iz g sear then p e rem ere O r ad perts in ation Ontari an ch p artn ainin BI co voca of cl er rog g tes. uding treatm s and ram ships 1/3 o ntribthe rese en s. nee and f the Bru arch ts is pro ers, urgen d to ac and no Gir vid p ce t. e o h ysici OB le stra a profe s, PhD ans, I has rate th te e the d, fo ssor in , a rese b ind r ustry rough pace o atri hippoca the firs the D archer rep t toget f c dis ep t m rese h ease pus. time, artmen at the ntati er Do s, su This the t Dop ves ch as disco role of Psy uglas , that men amine chia Men v sc er do hizo ta tr y in p tal illn is a neu phre opens pamin y at M l Healt cG e h nia. the ed w articula esses in rotran doo plays ill U Univ smit r ith r to in a niver ersity mem is that volve te r th a bet reg an d ory "Ou ion sity, h Instit ter and opamin imbalanat play r as u o te w u whic ork nder f the dem s lear e o stan bra nin is pre ce in th a centr lear h the helps g, an sen din in ca nal is go d th t in th chem role in apeu ning. In cause us bette f psy lled in at it ic e r tic chiplay hippo al. W brain fu app a few the bra under ha ro y ca s s i n

Sources: * 2015 Publishers Sworn Statement, ** LAB &E Bio Business 2011 Reader Survey, ***Publishers Sworn Statement, August 2016 Th e Ro le o f Do pam ine

in th eH ippo cam pus

3/15/19 4:05 PM


Deep Learning on the Move Montreal’s AI hub strengthens as MILA moves closer to other industry leaders BY KEREN STEPHENSON BIOLAB BUSINESS WINTER 18/ 19



ast fall, a specially formed logistics team at MILA– formerly the Montreal Institute for Learning Algorithms, now the Quebec Artificial Intelligence Institute – reviewed its plan to seamlessly shift the 20-year-old laboratory from the grounds of the University of Montreal to the O-Mile-Ex, a brand-new $60-million complex situated in Montreal’s Mile-Ex district. The organization comprises world-renowned researchers in the area of deep learning and machine learning. MILA’s chic new 100,000-square-foot office, constituting a quarter of the O-Mile-Ex building, is designed to generate an ecosystem where Montreal’s top researchers will together foster creativity by providing the

ultimate work/play environment, complete with pingpong and foosball tables, yoga and Pilates facilities, and its own catering service, The Mile-Ex district, now referred to as Montreal’s AI hub, has housed video-game developer Ubisoft for two decades, and more recently became home to Cortaix, IVADO, Element AI, Facebook and Imagia, who will share the O-Mile-Ex building with MILA. Meanwhile, Microsoft will soon occupy a nearby space. “Most of the important contributors to the Montreal AI hub will be in the vicinity,” says Associate Professor Aaron Courville, who joined MILA back in 2004 and is now considered one of the top AI researchers in Canada.


The main reason to become a non-profit was to concentrate as much of the core machine learning and AI talent in one spot, so we created this extra entity that allows us to bring people together from these institutions.

“I think it will create a really interesting ecosystem and make collaborating really easy for us.” Courville stresses that the timing for the relocation is increasingly important due to MILA’s rapid growth in recent years. Once a small research centre affiliated with the University of Montreal, MILA is now a worldrenowned independent non-profit organization known for its developments in deep learning, a type of programming that enables computers to learn from their experiences and develop knowledge of hierarchical concepts. Professor Yoshua Bengio, who is considered one of the fathers of deep learning, created the MILA lab at the University of Montreal in the 1990s with the ambition to further understand the principles of learning that yields intelligence. Bengio’s ambition to unite experts in the field led to MILA’s affiliation with McGill University in 2006, and in subsequent years, with the Polytechnique Montreal, and the HEC business school. The extended team now comprises more than 250 interns, students, staff members and professors, including Professor Courville. “The main reason to become a non-profit was to concentrate as much of the core machine learning and AI talent in one spot, so we created this extra entity that allows us to bring people together from these institutions,” Courville explains. “At McGill they are experts in reinforcement learning, or neural maps, and over at the University of Montreal we specialize in deep learning, so we have a combination of the two.” The combination of deep learning and reinforcement learning in one space is what makes MILA unique. The foundation is acclaimed for many breakthroughs in both fields, including the development of deep-learning algorithms that are applied in many domains, such as neural language modelling, object recognition and neural speech recognition. “There are really three hubs of AI innovation in Canada,” Courville says. “These are: Amii [Alberta Machine Intelligence Institute] in Edmonton, who are really well known for their reinforcement learning; the Vector Institute in Toronto, who are a very strong group in deep learning; and MILA in Montreal, where we have been fortunate to combine both.”

Top: University of Montreal Associate Professor Aaron Courville joined MILA in 2004, and is now considered one of the top AI researchers in Canada. Above: Margaux Luck is a bioinformatics PhD, applied research scientist, and member of the MILA Technological Transfer team. She’s solving concrete problems through research and technology transfer, and she’s training the private sector in deep-learning practices.


– Aaron Courville, Associate Professor



Courville explains that MILA’s mission has three pillars. The first is academic research: “From my point of view, that is what we come from and what we will always be: a worldclass institute, graduating top students who come from all around the world to work with us.” As part of this pillar, MILA runs summer schools in conjunction with Vector and AMII, to teach principles of deep learning and machine learning to graduate students from around the world. The second pillar, tech transfer, aims to stimulate economic growth by offering consulting services to companies who wish to integrate AI into their products and services. MILA’s team plays a pivotal role in AI training though its various corporate programs, as well as its Professional Master’s degree in Machine Learning. MILA’s final pillar is AI for social good – innovations that meaningfully impact society – such as applications in medicine, but also generative modelling for images, by studying large volumes of data and then training a model to generate similar data. Courville explains: “We look at how we generalize or have the ability to take the principles that we learn on, and generalize them to examples that we have never seen before, such as an autonomous vehicle in a new situation. The idea is to make these applications more robust and safer. We look particularly at language and how we can combine words together that we have never combined before in our life and they still make sense.” MILA currently runs almost 20 ongoing projects. Margaux



Luck, an applied research scientist at MILA, focuses on the medical applications of deep learning, mostly with startup businesses such as Imagia, with whom she develops models that help detect early disease and predict patient outcomes. “I’m also working on models to predict the arrival of a kidney in a statistical form, as well as predicting some of the characteristics of the recipient and the donor,” she says. Luck’s research aims to improve both decision-making in patient care and treatment, and survival rates of graft patients. On an average day, Luck may conduct research, write conference papers or teach doctors about machine learning. “I may also talk to the companies that I’m working with and advise them or help them to understand the model they are using. Our clients don’t necessarily understand machine learning and deep learning at the same level as me, so I spend time explaining how to go about analyzing,” she says.

MILA’s mission has three pillars: academic research, tech transfer and AI for social good.


MILA also conducts research sponsored by companies such as Microsoft and Facebook – both recent donors of approximately US$7 million – to improve learning algorithms for speech and language, and other areas of deep learning.

MILA’s chic new 100,000-square-foot office takes up a quarter of the O-Mile-Ex building. Designed to foster collaboration, creativity and discovery, the space encourages unencumbered thinking through work and play, complete with ping-pong and foosball tables, yoga and Pilates facilities, and a catering service.

of data with information and rules on a certain subject, such as language. This data is then used to train the models to understand speech and image recognition and make inferences and decisions based on this information. “GPUs are by far the most important pieces of equipment we use, and we need a lot of them,” Courville explains. “Deep learning and reinforcement learning these days are both extremely computationally intensive. These models are quite large, and they need a lot of data to train them. And all of that translates into a lot of hours of computing on these machines.” MILA requires large quantities of the latest GPUs to stay ahead in the field: “We are still an academic lab and we are partnering, but also competing with, industrial labs who have much greater computational resources than we have, and keeping up with them in that direction has been a challenge for us,” says Courville. As MILA continues to champion its deep learning and machine learning techniques, its researchers provide a platform for companies to improve practical applications for these techniques, such as self-driving cars, improved voicerecognition and automation of business and medical decisions. Because MILA remains fundamentally part of Montreal’s universities, however, its researchers and students are free to choose their subject of research. The advantage to MILA’s partners is their free access to this research in exchange for funding, as well as privileged access to employ graduates from a world-renowned AI program.


Luck also refers to the collaborative environment at MILA, which provides a stimulating backdrop for creativity – an aspect that fits MILA’s primary pillar of academic research and collaboration.“There’s a lot of diversity at MILA,” says Luck. “I mean, we still have more men than women who work here, but that’s kind of normal in this field. But we are from a lot of different backgrounds such as physics, biology, mathematics and computer science, and we have students from all around the world who come here. There are a lot of smart people,” she adds. “We have a great community.” MILA also conducts research sponsored by companies such as Microsoft and Facebook – both recent donors of approximately US$7 million – to improve learning algorithms for speech and language, and other areas of deep learning. But as a non-profit, the key objective is not to make money; MILA also consults smaller companies to promote learning in the AI space. “These are fairly small companies that would not necessarily be able to go to some of the other opportunities, so this is part of our contribution to the tech community,” Courville says. MILA will receive up to $80 million in funding from Quebec over the next five years, as well as $44 million from the Canadian Institute for Advanced Research (CIFAR), an organization that coordinates Pan-Canadian AI strategies to ensure that Canada remains a world leader in the field. “They [CIFAR] are a pretty key aspect of what we do around here,” Courville admits. While MILA’s research is complex, its lab equipment remains fairly simple – most work relies on GPUs (graphical processor units) that collect large volumes



NEW: Multichannel pipettes for 384-well plates

384. Ready. Set. Pipette! Manual pipetting of 384-well plates made easy Do you work with 384-well plates and strive for more efficient manual pipetting methods? Your worries are over. The new Eppendorf Research® plus and Eppendorf Xplorer® plus 16- and 24-channel pipettes and pipette tips make working with 384-well plates convenient, ergonomic and safe. Feel the magic of filling a 384-well plate within a minute.

> Reduce mistakes and the strain of filling a 384-well plate with single-, 8-, or 12-channel pipettes > Set up or validate an automated process manually, and have a backup system in case automation fails > Easily scale up from 96- to 384-well plates to increase throughput • 800-263-8715 001.A1.0153.A Eppendorf®, the Eppendorf Brand Design, Eppendorf Research®, and Eppendorf Xplorer® are registered trademarks of Eppendorf AG, Germany. All rights reserved including graphics and images. Copyright © 2019 by Eppendorf AG.













he human brain has about 86 billion neurons interconnected with 100 trillion synapses that shoot messages to each other in thousands of configurations. These are big numbers and big data in an organ that takes up only two per cent of your entire body weight, while continuously performing complex neural computations – making for an impressive and mysterious cognitive machine. Now, a new state-of-the-art computer platform at Toronto’s Centre for Addiction and Mental Health (CAMH) will bring neuroscientists closer to unravelling those very same mysteries, as they work to resolve brain disorders. The system harnesses massive amounts of data from medical imaging, genomics and clinical research. Late last summer, teams of scientists and systems engineers from Dell EMC and the Krembil Centre for Neuroinformatics at CAMH went live with what can only be described as a behemoth of computation, the CAMH Neuroinformatics Platform. Like the Ontario Brain Institute’s Brain-CODE platform, CAMH’s new tool is a centralized research database that secures, manages and organizes complex data for researchers from within the organization and around the world. David Rotenberg, Krembil’s Operations Director, says that the idea for launching the million-dollar initiative, which took about a year and a half to develop, was to have a robust, centralized research data management superstructure, which he sees as a critical piece missing in many larger institutions. “At CAMH, our goal is to understand how the human brain works in health and in disease; and in particular, to understand what the biological underpinnings of these disorders are,” Rotenberg At CAMH, our goal is to explains. “A lot of the stigma associated with understand how the human brain mental illness is that it’s works in health and in disease; something mysterious; it’s something in the and in particular, to understand mind. Whereas with what the biological underpinnings other disorders, such as of these disorders are. cancer, there are clear biological causes that people understand; and, – David Rotenberg, Operations Director, Krembil when you understand the cause, it helps to remove that stigma, but it also helps us move towards treatment. “Mental health is something that can be understood and it can be solved. By collecting these different types of data, we can look across the scale from the gene to the brain cell to the small circuits, and then to the overall structure of the brain, and then to how that alters the symptoms that we see in the clinic. Data is at the heart of being able to understand the relationship between the hierarchies and where these issues might lie, but also the full spectrum of diagnoses and how they may overlap.” With 50 research projects well underway since the platform went live, it clearly answers an important call. The system currently holds data from 34,000 researchers and links to 380,000 CAMH patient records, for a whopping 15 terabytes of datasets.



research projects well underway since the platform went live, it clearly answers an important call.

The system currently holds data from

34,000 researchers and links to

380,000 CAMH patient records, for a whopping


terabytes of datasets.


With so many records, cyber and data security are critically important for CAMH and its administration. Participating researchers must sign several stringent agreements before joining, and they fall under the jurisdiction of the CAMH’s Research Ethics Board. Server rooms are under lock and key, and only certain people have access. All systems are password protected and sit behind multiple firewalls, constantly scanned for security breaches, intrusions or viruses. To protect client confidentiality, patient names and addresses are kept separate from their corresponding data. “You cannot combine the two, so if you had someone’s brain scan you wouldn’t know who they are,” Rotenberg says. The system organizes data in advance, saving researchers from spending weeks and months “digging” to find, organize and make manual links to other datasets. “There are lots of different groups of researchers with different data needs, such as MRI, PET, EEG, epi-genetics, genetics,” Rotenberg explains, “all of which are supported under the same framework.” In Canada, CAMH is a member of the Canadian Open Neuroscience Platform, which brings together many of the country’s leading scientists in an interactive network of collaboration. It’s something that Rotenberg says will ensure the interoperability of data, something which he sees as vitally important. “Interoperability is a key principle toward bringing institutions together in meeting these challenges rather than being isolated to ask those research questions without the larger data.” Globally, CAMH works with Europe’s Human Brain Project and Switzerland’s Blue Brain Project. It’s a “natural fit,” according to Rotenberg, because there is mutual interest in understanding the human brain at the cell level and the whole brain level. For new researchers accessing the system, CAMH offers workshops twice a month and quarterly. “We’ve seen trends that show you need to know computation in this day and age, to be able to ask the right questions when faced with these huge volumes of data,” Rotenberg notes. “We train folks on accessing the system, how to use the visualizations and to support data analysis.” Researchers easily can transfer their data into the platform and integrate it with data from other sources; as Rotenberg says, “This really allows people to share and interact in the same consistent environment no matter where they’re from.” All the data are secured by permissions for specific research groups. By using the Oracle Virtual Machine (OVM), CAMH adopted a primarily

In Canada, CAMH is a member of the Canadian Open Neuroscience Platform, which brings together many of the country’s leading scientists in an interactive network of collaboration.


A behemoth of computation, the CAMH Neuroinformatics Platform is capable of harnessing massive amounts of data from medical imaging, genomics and clinical research.

virtualized architecture in which one computer performs like many, or “virtual” computers. In the CAMH platform, these virtual machines contain the operating system and the software, but are not tied directly to the physical machine. The OVM enables flexible deployment, efficient snapshots for backup and simplified fail-over procedures to switch on virtual machines, particularly if hardware were to fail. “Part of our backup strategy is that we have half of these arrays at one site and the other at the other site of our two campuses,” Rotenberg explains. “We use that for backup and fail-over, so if one site were to go completely down we can use the other site to access all the data. The actual data is backed up every hour on all these systems. We basically take snapshots of the data frequently so if the data were deleted accidentally, we can roll back up to a month.” To store all the data and backup functions across the two sites, it takes 1.9 petabytes of memory. When you consider that the average mobile phone has about 16 gigabytes, that’s about the same memory you might find if you combined 125,000 mobile phones. These data are priceless, says Rotenberg, adding that it costs about $1,000 for a typical genetic sequence and as much as $10,000 for some of the newer scans, “so we take very precious care of our data, and I get to sleep at night.”



Rethinking THE LAB



Alán Aspuru-Guzik is igniting a revolution with AI and robotics – and maybe, saving the world


Automation is not new to chemistry. While it’s been around since the ’90s, Aspuru-Guzik’s work has taken it even further, by adding “a brain to it.”


eet Bob, the Bayesian Optimized Bartender. He can fix a great Tequila Sunrise. Bob’s smart enough to know just the right amount of orange juice to mix with grenadine and tequila. If you don’t like it, Bob will make it better next time, and even recommend another drink altogether based on your preferences. Bob was developed by a team at the University of Toronto’s Matter Lab led by computational chemist Alán Aspuru-Guzik, as a small demonstration of the types of powerful programming and integrated tools that can open the gate to self-driven laboratories, where discovery is accelerated. “If you look at the chemistry lab of the 16th century or even the 21st century, you will see the same thing in terms of how chemists run experiments. Nothing has changed. If we really want to rethink discovery, we need to rethink the laboratory,” said Aspuru-Guzik last year at a meeting of the American Chemical Society. The former Harvard professor, born in the U.S. but raised in Mexico, has devoted his academic life to finding ways to simulate matter more effectively, seeking answers to the problems posed by molecular design, through quantum chemistry and artificial intelligence. It’s a world where chemistry, physics, applied math and computer science meet. After the last U.S. election, Aspuru-Guzik decided to continue his research in Canada, a nation he says is leading the world in artificial intelligence and quantum computing. Fittingly, his first day on the job in the departments of chemistry and computer science at the University of Toronto was Canada Day 2018 – a day of celebration. One of the other things that attracted Aspuru-Guzik to Canada was the offer of receiving a Canada 150 Research Chair in Theoretical and Quantum Chemistry. He is also a faculty member for the Vector Institute,

an artificial intelligence think-tank headquartered in Toronto. On his arrival, Aspuru-Guzik launched the Matter Lab, comprising a multidisciplinary team of chemists, physicists, computer scientists and others working on applying algorithms to simulate chemical molecules and understand their interactions. Designing effective chemistry experiments is complicated and takes a great deal of time and resources. “We do great experimentation now. Experimentalists usually take pieces of the puzzle one by one and have not fully embraced automation, at least in chemistry,” AspuruGuzik explains. It’s a bottleneck that slows down muchneeded new discoveries. “I wanted to accelerate time to discovery. On the global scale, I’m really worried about all the problems that humanity faces. Although science and technology have provided many solutions, there are many more to be provided,” he reflects. “Therefore, it is only fair that we scientists, who are using all this public money, try to make ourselves more efficient in finding solutions that society trusts us to solve.” Automation is not new to chemistry. While it’s been around since the ’90s, Aspuru-Guzik’s work has taken it even further, by adding “a brain to it.” “We are going to have artificial intelligence drive every decision the autonomous system is making in such a way that is as near to optimal as we can in decision making,” he says. This offers greater precision at resolving chemical problems. “Done together with a human that is co-guiding this machinery, we are able to accelerate the process.” Along the path to building a fully automated laboratory, you need quantum computers that can accurately simulate molecules – machines that are millions of times more efficient than today’s computers. The problem is that only a handful of these devices exist now. The notion of molecular simulations stems from the work of Nobel-prize winning physicist Richard Feynman, who in 1982 found a way of thinking about quantum simulation in a way that was numerically exact. This led to a paper in 1997 by Seth Lloyd at MIT, where he showed how





to actually do it. “In 2004, I looked at that literature to find if it was applicable to chemistry, and then began to develop more and more algorithms,” says AspuruGuzik. To run these algorithms, early-scale quantum computers are currently being used by people around the world, including Google, IBM and his own startup on the commercial side, Zapata Computing. It fits in with his group’s longer-term goal to be prepared so that “by the time the quantum computer is large enough, we’ll have an arsenal of tools to address problems of a chemical nature.” This global enterprise of people trying to solve these – Alán Aspuru-Guzik chemistry problems is looking to quantum computers as one of the earliest possible applications. “It’s been a great adventure being a part of this amazing group of people thinking about this,” he says, but the timing of it depends on the hardware and newer and better algorithms. At the Matter Lab, the team uses such algorithms in combination with robotics to create what they call “materials acceleration platforms” (MAPs), or “self-driving laboratories.” In collaboration with scientists at UBC, the team devised a way to orchestrate and synchronize automated instrumentations, manage databases and interact with various artificial intelligence algorithms. It’s called ChemOS, an orchestration software that supplies the structure for operating autonomous laboratories. Using machine learning to interpret commands from users, it then plans, executes, evaluates and refines experiments. It can also interact with other automation software like the programs that control robotic arms or solution injectors on laboratory machines. Aspuru-Guzik compares ChemOS to a brain, performing the higher-level functions needed to coordinate a successful experiment: “We believe that eventually, every piece of equipment in a lab will be able to talk to a network and fit in with the Internet of Things.” Remember Bob, the robot? Through ChemOS programming, Bob figured out how to mix the perfect Tequila Sunrise by controlling robotic liquid pumps and selecting from a variety of ingredients, including vinegar. With feedback from humans, the program learned to avoid the vinegar to get the cocktail just right. While the exercise helped to demonstrate autonomous solution mixing, it also showed how human input could be important. Aspuru-Guzik acknowledges there’s great work undertaken by individual labs around the world, but he’d like to see more collaboration. “We have to remove our egos and bring our Bob, the Bayesian Optimized Bartender, superpowers together in a larger collective like the League of developed by U of T’s Matter Lab, has learned to mix the perfect cocktail. Justice. We don’t have time not to,” he warns.

We believe that eventually, every piece of equipment in a lab will be able to talk to a network and fit in with the Internet of Things.



WINTER 18/19

» The science of food and beverage

Canada’s Food Guide taps into the “protein wars”


growing numbers of people have reduced their consumption of meat and milk products, the Canadian government also has reevaluated past definitions of a “balanced” diet, as we see in the dramatically revised Canada’s Food Guide. With big changes, however, come new questions: How do we define “healthy” (as in: “eat a variety of healthy foods each day”)? Are we steering consumers in the direction of those foods marketed as “healthy,” without informing them how to differentiate between facts and “spin”? With advice such as “Take time to eat,” are we oversimplifying the complex issues of food-on-the-go and a decrease in cooking skills and family dining? One line of the new Canada’s Food Guide caught extra attention: “Choose protein foods that come from plants more often.” As Dalhousie University discovered in its 2018 survey, Plant-based dieting and meat attachment: Protein wars and the changing Canadian consumer, a significant proportion of respondents (17.4 per cent) identify as following a variation of meatless, or reduced meat, diets. Among the vegetarians (2.1 per cent) and vegans (1.1 per cent), a larger group arose, representing 10.2 per cent of the 1,027 respondents: “flexitarians,” who eat meat and fish “occasionally.” Less than half of respondents admitted to eating meat on a daily basis, while a surprising 11.2 per cent indulged rarely (once or twice a month, at most). Critics of the revised Food Guide quickly pointed to the challenges faced by low-income families – “make healthy choices” is easier with more money. Canada’s Food Price Report 2019 (a collaboration between Dalhousie University and the University of Guelph) estimates the average Canadian family is expected to spend $12,157 on food this year, while the cost of fruits and vegetables is likely rising by 4 to 6 per cent. When you consider that more than one-quarter of Canadians make less than $20,000 per year, according to the most recent census, the cost of food is a hefty expenditure for many families; nationwide, the median income overall was $34,204 in 2015 (men: $40,782; women: $28,860). If you look at the apparent gender/income differences, that basically means the average woman can’t afford healthy food. (Let’s not talk about the cost of housing!) These conversations arise when the government encourages us to “eat less processed foods” and “choose healthier menu options.” In many cases, a family is lucky to eat at all. As we consider the larger topics of sustainability, food insecurity and waste, there are multilayered issues that require thoughtful discussion and serious solutions. You can lead a horse to water… but what if there isn’t any potable water?


We’re excited to announce that Canadian Food Business now will be included in every issue of BioLab Business, with expanded 16-page sections in spring (March/April) and summer (July/August).

FALL 2018

» The science

of food and beve


FALL 2017

» The science of food

High-tech Food Safety

How Blockchain and DNA analysis can change the game



AbsolutLY ENGAGING The vodka giant taps into

Putting the spotl ight on local

and beverage

Canadian cuisin e

Toronto pride


Labels: Keep it Clean!


New technology transforms manufacturing

Are BreAd SAleS Turning STAle?

New s Bites Dairy industry challenges Canada’s Food Guide Dairy Farmers of Canada (DFC) released a statement shortly before the new Food Guide was released, stating that the elimination of the Milk and Alternatives group, while encouraging the consumption of plant-based protein, could be “detrimental to the long-term health of future generations by leading them to erroneously think that dairy products are unhealthy.” In light of recent trade agreements, this threatens an industry that is already struggling. “There is no scientific justification to minimize the role of milk products in a healthy diet, as they are a key source of six of the eight nutrients that most Canadians already fall short of,” said Isabelle Neiderer, DFC's Director of Nutrition and Research, and a registered dietitian. “The current scientific evidence clearly demonstrates that the daily consumption of two to four servings of milk products has a beneficial role to play in promoting bone health and preventing several chronic diseases that Health Canada wants to address with the new Food Guide, such as hypertension, colorectal cancer, type 2 diabetes and stroke.” For more than 75 years, milk and dairy products have been clearly recognized within Canada’s Food Guide as playing a key role in a healthy, balanced diet. In 2013, Statistics Canada released the results of a study involving 6,400 individuals aged 3 to 79 years old, which determined that 32 per cent of Canadians have insufficient vitamin D levels, while those who consumed milk at least once per day had a higher average, and more of them were above the vitamin D cut-off for optimal bone health.



Plant scientists ‘hack’ photosynthesis A team of researchers at the University of Illinois developed genetically engineered tobacco plants capable of growing faster and up to 40 per cent bigger; the scientists inserted genes to simplify and speed up the detoxification process (photorespiration) that occurs when plants process oxygen instead of carbon dioxide. RuBisCo, the enzyme in plants that captures carbon dioxide from the air, is not always capable of distinguishing the difference between the elements. The plant then expends energy “detoxifying” itself that could otherwise be used for growth. As the researchers note in their study (published in Science), “Photorespiration reduces the photosynthetic conversion efficiency of light into biomass by 20 to 50 percent, with the largest losses occurring in hot dry climates where yield increases are sorely needed.” The team is now testing the same technique on potatoes, soybeans and cowpeas.

Consumer dining habits shift – and restaurants celebrate In a survey of 2,000 Canadians, 71 per cent dine outside the home up to 10 times per month, according to Eagle Eye’s recent Changing Tastes & Flavours report. Notably, younger generations and working professionals dine out more frequently, up to 20 times per month. For these meals, almost 40 per cent of respondents spend $30 or more, and 13 per cent use third-party delivery services such as Uber Eats.


dine outside the home up to

10 times a month Canadian milk industry shows improvement A 2018 study conducted by Groupe AGECO evaluated the environmental performance of the milk industry in 2016 in comparison to data from 2011, finding that the quantity of milk produced per cow has increased by 13 per cent, while there were decreases in carbon footprint (7 per cent), water consumption (6 per cent) and land use associated with milk production (11 per cent). Milk produced in Canada has a lower carbon footprint than average; a litre produced in Canada emits 0.94 kg CO2 eq, which is about 1/3 the greenhouse gas (GHG) emissions compared to the global average.

Can AI feed the world?

Researchers at Penn State recently explored the possibility of using the natural antioxidants found in grain bran – alkylresorcinols (ARs) – to replace synthetic food preservatives. Plants such as wheat, rye and barley produce ARs naturally to prevent mould, bacteria and other organisms from growing on the grain kernels. ARs can also help protect against cancer, according to a review published in European Food Research and Technology. Usually the food industry discards the bran layer of cereal plants, or uses it for animal feed. “We're taking something that's usually discarded in a waste stream and turning it into something useful,” explains Andrew S. Elder, doctoral candidate in food science. “Currently, there's a big push within the food industry to replace synthetic ingredients with natural alternatives, and this is being driven by consumers. Our work is focused on identifying new natural antioxidants to extend the shelf life of food and meet consumer demands.” The team developed a technique to extract and purify ARs from rye bran, then studied how well ARs were able to preserve omega-3-rich oils in emulsions, where two fluids do not fully mix – for example, vinegar and oil. The researchers found that ARs did act as antioxidants in an emulsion, preventing omega-3 oils from spoiling as rapidly as they did in emulsions with no antioxidants added; however, ARs were not as effective as either alpha-tocopherol (Vitamin E), a natural antioxidant, or butylated hydroxytoluene, a synthetic antioxidant. The researchers noted that their AR extracts were not completely pure, which could have reduced the effectiveness of the ARs. Also, the researchers used a blend of different ARs that had different molecular structures. Future work looking at different types of ARs will reveal whether an individual AR type is more or less effective than conventionally-used antioxidants. Their findings were published in Food Chemistry vol. 272.

Plants such as wheat, rye and barley produce ARs naturally to prevent mould, bacteria and other organisms from growing on the grain kernels.

Food cruises around the world

Holland America Line now offers nearly 100 food tours around the world developed in partnership with Food & Wine magazine. Guests take cooking classes with top chefs, go on culinary walking tours, visit wineries, sample authentic street food and more; the tours are offered in Asia, Australia/New Zealand, the Mediterranean, Northern Europe, New England, Tahiti, Hawaii and (yes) Canada. Over the next year, tours will be expanded to include the Caribbean, Mexico, South America, Alaska and additional European ports.


Bran antioxidant keeps food fresh longer

Olam International, a leading global agribusiness operating in 66 countries, has partnered with Canvass Analytics (based in Toronto, San Francisco and Basel, Switzerland) to utilize artificial intelligence as part of its strategy to find innovative ways to meet the world’s growing demand for food, feed, fibre and fuel. “The insights offered by data and analytics provide opportunities to transform the way we operate and to find new ways to address challenges across the agricultural value chains,” notes KC Suresh, Managing Director and CEO at Olam International. “By integrating emerging technologies into our business, we are driving greater efficiencies, enhancing the sustainability of our supply chains and offering more value to our customers.” Funded by Google’s AI-fund, Gradient Ventures, Canvass Analytics provides an automated, AI-powered platform that helps its customers improve quality, lower production costs and reduce their energy consumption.




WINTER 18/ 19


By Jana Manolakos

Arlene Dickinson’s incubator helps small companies do big things a marketing mogul and a straight-forward, no-nonsense venture capitalist whose steely intelligence wins bids for startup companies on the long-running television series, Dragon’s Den. Although Arlene Dickinson gained notoriety as a “Dragon,” she’d prefer to keep that profile separate from one of her latest enterprises, District Ventures, a national accelerator for the food and wellness industries. Launched in 2015, it’s grown into a community for driven entrepreneurs where they can network, access strategic investors and receive the business development support they need to take their business to the next level. As CEO of Venture Communications and general partner for District Ventures Capital, a fund that supports the accelerator, Dickinson explains, “People might be attracted to my fund and to the accelerator because I am a part of it, but I’m not critical to it. What I’ve done is I have led the charge on it; I’ve created the infrastructure and ecosystem for it, but I haven’t done it alone. I’ve done it with a really incredibly strong team.” That ecosystem connects early-stage companies in the accelerator program with investors whose money is deployed by District Ventures Capital. Dickinson says those new companies also have access to mentors, sponsors, the IBM Innovation Space and Venture Communications’ marketing bench strength. “It’s fairly broad as it relates to the available resources, but at the heart of it, this is about deploying capital intelligently to help these businesses,” she says. The accelerator recently announced its sixth cohort. Among the seven fledgling enterprises, Jesse Clark, CEO of Maze Bar Inc., started his energy bar company as a teenager, funding it by working multiple part-time jobs throughout high school. Since closing an investment deal with District Ventures last August, he took part in their five-month cohort accelerator program and is now set to present to future investors. “It provided an overwhelming amount of resources and support to help take my business to the next level,” Clark says. “I’ve been connected with retailers, distributors and investors from across Canada, and have also learned more about the logistics of running and managing a business in five months than I did in two years of business school before dropping out to pursue Maze Bar full-time.” With support from District Ventures, Maze Bar is on pace this year to achieve five times the revenue it generated in 2018. As a proud Canadian, Dickinson believes there is a lot of expertise and opportunity in this country to do meaningful things. But, she sees a gap in the way that new Canadian companies who grow ingredients and innovate are supported in their development – particularly, in the agriculture, food and health sectors.




Canada exports a large portion of its raw materials to other countries, bypassing the opportunity to manufacture and sell its own products. “We are very good at production, but we need to then be able to take those ingredients, turn them into products and sell them to the world,” Dickinson says. “I am a big believer that no product should ever leave this country without value being added to Canada.”

A personal mission

Since its launch, thousands of startups have applied to join the accelerator, which to date has helped 60 companies from 21 Canadian cities generate over $65 million in revenue, raise over $12 million in capital and employ over 250 people.

Why the focus on the food and wellness sectors? She says that a new generation of consumers value healthier products, and are keenly focused on where their food comes from, how it tastes, its safety and security. “As I age, I’m thinking more about what I consume, how I stay healthy and fit, and the impact of what I eat on my ability to age well and gracefully,” Dickinson admits, explaining that the convergence of her personal journey with what was happening in the marketplace led to the accelerator’s focus on food and wellness. What sets District Ventures apart from most incubators and accelerators in Canada is its unique position as a company that’s unattached to post-secondary institutions or research facilities; instead, it’s run by a successful entrepreneur known for her expertise at commercialization and how to get products into the hands of consumers. Since its launch, thousands of startups have applied to join the accelerator, which to date has helped 60 companies from 21 Canadian cities generate over $65 million in revenue, raise over $12 million in capital and employ over 250 people. Initially, District Ventures did not provide funding, but instead connected members to capital by introducing them to potential finance sources. In 2017, the company announced a



change in the model, with the first two cohorts in 2018 receiving a $130,000 investment in exchange for a minority equity stake in their companies. “We learn with every cohort, what we can do more effectively, how we can support differently,” Dickinson says. The opportunity for early investment along with growth programming offered by the new model is expected to increase interest and demand. Despite statistics that show most startups fail within the first three to five years of operation, District Ventures’ earliest cohorts seem to be doing just fine. “This last group has actually been really impressive in terms of the number of jobs that have been created and the capital that has been raised,” she says. Dickinson admits some have caught traction more quickly than others, but she’s quick to point out that the time it takes to scale can vary depending on the business. “We’ve seen some really successful companies come out of there – Bow Valley BBQ, Chickapea Pasta and Drizzle Honey – I could name a dozen that have done incredibly well. I don’t believe there are any that haven’t been able to maintain a business up to now.” For gourmet sauce company Bow Valley BBQ, it’s been an astounding success story. Company founder Jamie Ayles worked in restaurants most of his life – he started at age 12 and moved his way up in kitchens around the world, while winning multiple culinary awards. In 2012, things took an interesting turn when he and a colleague used his personal recipes to launch Bow Valley BBQ. It was a good start, and it became even better the following year, when he purchased a local salad dressing company called Boccalino Fine Foods. “What I’ve been able to do with Boccalino




“Yes, AI is a different path that changes how we learn and operate, and it’s a threat, but everybody thought the steam engine was also a threat. People will just need to adapt and change with it. We can’t immune ourselves to the threat, ignore it or hope it goes away. The truly successful people of the future will understand this and reapply themselves in new ways and learn new skills. The next generation will think very differently about what their work life will be.”


in the last two years is become the largest independent Canadian-owned salad dressing company in Canada,” he explains. And if things were good then, they got even better shortly after, when he was accepted into District Ventures’ second cohort, three years ago. “Over the past two years, we’ve done a partnership deal with District Capital, and since then things have really taken off for us,” Ayles says. “At the time, we already had good momentum and a good trajectory, but we needed the resources and the networking to manage that successfully. In the last few years, our business has grown about 2,500 per cent.” Ayles appreciated the straightforward advice Dickinson gave, working directly with her and her team, particularly with fund manager Jason Berenstein. “The cool thing about working with Arlene is that she has access to a ton of outside networking resources and mentorships,” he says. Some of the people Ayles met included Dani Reiss, CEO of Canada Goose, and Larry Tanenbaum, chair of Maple Leaf Sports & Entertainment. As mentors go, he’s learned from people who have experienced both success and failure. “I think it’s equally important to learn what not to do as well as what to emulate,” he admits. For Dickinson, the secret to success isn’t clear cut: “A good entrepreneur is constantly pushing themselves and learning what they don’t know, and surrounding themselves with people who can assist them. A good entrepreneur knows when they are standing in the way of business. These are all things that happen with time and effort. People will fail and make mistakes, and that’s part of the learning and the opportunity.” District Ventures is a project Dickinson is wholly vested in, both personally and financially; it’s something she would love to have as a legacy. “We are building a very unique ecosystem in this country, everything from supporting with capital to helping earlier-stage startups to helping more early- to midstage companies grow and adding technology,” she explains. “We are doing something that hasn’t been done in Canada. We believe that food and health and taking Canadian expertise to market is really important, and we appreciate the support we are getting.”

AS DICKINSON NOTES, “JOBS ARE CHANGING AND THE WORLD IS EVOLVING. What has not changed is the innovative thinking of humans and the ability to create and understand how to build meaningful organizations. It’s not something that artificial intelligence will resolve. “AI can think about things and create new ways of doing things, but I personally believe that humans and our ability to problem-solve from a practical perspective, in building business, is in itself unique. You are seeing more ‘solo-preneurs,’ and while I’m not a big fan of any kind of label, you are seeing more individuals going out and running businesses. You are seeing more people doing things outside of their day-to-day jobs, creating other streams of revenue and new ways of doing things. You are seeing technology applied because humans are learning how to apply it more effectively.







Across International just released its newest magnetic stirrer, which spins 33 per cent faster than the previous model, and can stir up to one gallon of liquid. A maximum speed of 2,000 rpm makes it one of the speediest on the market. Ideally suited for stirring and homogenization, this unit works particularly well with high-viscosity liquids. The heated, ceramic-coated aluminum plate is longer, allowing it to hold a 7-inch beaker perfectly. It can heat up to a maximum temperature of 350°C, much hotter than the industry standard. A dual thermocouple system allows for more accurate readings, and a digital keypad makes it easy to program.

In February, Labnet International introduced its next generation of gel documentation systems, the ENDURO™ GDS II and GDS Touch II. Used for visualization of DNA and protein stained within a gel, the GDS II comes in either a 302nm or 365nm version and runs on Windows-based PCs. The GDS Touch II comes in either a 302nm or 365nm wavelength version, with 470nm epi-blue lights for use with Safe Dyes that eliminate the need for ethium bromide; it also includes a builtin Windows 10 tablet computer. Photos can be filed as jpeg or 16bit Tiff files, ideal for publication.



Fujifilm Medical Systems USA debuted the latest version of its Synapse 3D Advanced Visualization software recently, expanding its 50 clinical applications with five new ones including: Breast Analyzer MR, Delayed Enhancement MR, Endoscopic Simulator, Intravoxel Incoherent Motion (IVIM) MR and Prostate Viewer MR. The company’s Vice President of Medical Informatics, Bill Lacy says, “Packed with new applications spanning multiple specialty areas, our latest version was designed to improve workflows and diagnostic confidence while enhancing comprehensive and collaborative care.” The new version was designed with extensive clinical input and incorporates intuitive, logical workflows that guide users to efficient and accurate results.

Histological sections of such things as the blood vessels of a cat’s eye spring to life with greater image contrast using the Leica Microsystems SP8 FALCON, an award-winning innovation in confocal imaging. It combines confocal and Fluorescence Lifetime IMaging (FLIM) capabilities to deepen exploration of sub-cellular processes. The SP8 FALCON enables scientists to utilize the recognized power of FLIM for functional imaging without needing to add complex equipment to an existing confocal system. Visualization and tracking of dynamic subcellular processes becomes fast and simple. It also allows FLIM to be combined with other imaging techniques offered with the SP8 confocal platform, such as multiphoton excitation or stimulated emission depletion.


AUTOMATION SOFTWARE HELPS LAB TEAMS DO MORE WITH LESS Green Button Go™ 2019 is Biosero’s scheduling software for integrated laboratory workcells. Designed to be easy for users, it can control simple to complex workstations using the most advanced software technology available. The simplified drag-anddrop user interface enables more advanced scheduling and operating features, such as input parameter screen development, error logging and external notification. Workcell programming is more convenient using the drag-and-drop instrument/ device control commands accessible from the scheduling main screen. Among many features, it offers real-time feedback at a glance, multiple user accounts, simulations for testing workcell, and functionality that simplifies assays with complicated requirements.

“WIPE-ON GLASS” PROTECTS FROM INFECTION THROUGH ANTIMICROBIAL SURFACES Nano-Care Deutschland AG has launched a third generation of ultra-thin coatings it calls “wipe-on glass,” which disinfect any surfaces for several months. The product creates a robust barrier to pathogen microorganisms such as MRSA, mould, fungi and yeast. After curing, the surface is slightly rough on the microscopic scale, thus destroying the cell wall of the colonizing microorganism, without leaving perceptible traces. Wipe application is simple and consumer-friendly for a wide array of objects, from smartphones to toilet lids. The company says its self-disinfecting wipes represent the next step in self-cleaning coatings.

NEW ANALYTIC TOOL HELPS ADVANCE GENOMICS AND MOLECULAR BIOLOGY RESEARCH The Femto Pulse System is ideal for genomics and molecular biology research, providing laboratories with true versatility in sample analysis. It can separate large DNA fragments and smears, BAC clones and total RNA with speed and can accurately quantify, qualify and size DNA and RNA samples. Perfect for low-concentration and/or large-size nucleic acid samples, it easily analyzes diverse samples including: cfDNA, total RNA, genomic DNA, large fragment DNA, messenger RNA (mRNA) and more. DNA and RNA samples are effortlessly separated on the same capillary array, and two different gel matrices can be loaded, enabling the unattended, sequential separation of RNA and DNA samples without cumbersome array swaps or cleaning.

A revolutionary new addition to Epson’s 6-Axis robot lineup, the Flexion N2 stands tall in a class of its own. Featuring patented compact folding arm technology, this innovative robot offers significant advantages in efficiency of motion and workcell space reduction. With the ability to operate in a tight space, the Flexion N2 can be utilized in laboratories where traditional 6-Axis robots cannot. Designed with a smart new kinematic configuration, the Flexion N2 has a 450mm reach and a 2.5kg maximum payload. Epson’s unique Residual Vibration Technology allows arms to produce outstanding speed, precision, accel/decel rates and much more, while significantly reducing vibration.

LIST OF ADVERTISERS & WEBSITES VWR Page 2................................................................................. EPPENDORF Page 4 CANADIAN PHARMACEUTICAL DISTRIBUTION NETWORK Page 16................................................................................. BIOLAB CANADIAN FOOD BUSINESS Page 25................................................................... EPPENDORF Page SENSOR HEALTH Page VWR Page 48...............................................................................









rtificial intelligence is not a new idea – it’s been around for hundreds of years. The ancient Greeks touched on it in their myths. Jonathan Swift conceived it in his 1726 description of “The Engine” in Gulliver’s Travels. Today, it’s no longer the stuff of fantasy, but a burgeoning area of research with implications for everything from better medical diagnoses to self-driving cars. In fact, the market for AI-related products is predicted to reach $47 billion by 2020, according to the International Data Corporation (IDC), and the field has attracted significant investment from Google, Facebook, Baidu and other technology leaders. In 2017, the Canadian government jumped on board with $125 million in funding for a Pan-Canadian Artificial Intelligence Strategy, handing the playbook to the Canadian Institute for Advanced Research (CIFAR), a renowned hub of scientific inquiry founded in 1982 by Fraser Mustard. CIFAR has since been working with three new AI institutes for deep learning and reinforcement learning: the Alberta Machine Intelligence Institute (Amii) in Edmonton, the Montreal Institute for Learning Algorithms (MILA) and the Vector Institute in Toronto. CIFAR also funds AI and society programs, and this past December organized the first annual meeting of the Pan-Canadian AI Strategy, which showcased the first cohort of Canada CIFAR AI chairs, and 40 of the nation’s leading AI researchers. Canada’s emergence as a global AI leader is due in large part to the decades-long work of a group of researchers funded by CIFAR, led by Chief Scientific Advisor at the Vector Institute, Geoffrey Hinton (also with the University of Toronto), along with Yoshua Bengio at the University of Montreal, who launched MILA, and Yann LeCun, Facebook’s Director of AI Research. Their work made fundamental advances in AI, driving knowledge that enabled computers to better perceive patterns and make accurate predictions using artificial neural networks, similar to how humans learn. Based at 133 universities and research centres around the world, CIFAR has 404 fellows, scholars and advisors. Since its inception, it has been associated with 19 Nobel Laureates. In addition to funding from the governments of Canada, British Columbia, Alberta, Ontario and Quebec, CIFAR also receives support from national and international partners, as well as individuals, foundations and corporations. CIFAR’s international partners include the Gordon and Betty Moore Foundation, Bill and Melinda Gates Foundation, Chinese Academy of Sciences, and Inria in France.



Thermo ScientificTM Capitol VialTM food grade polypropylene, triple seal design containers, with flip-top lids are aseptic, airtight, and leak resistant, and are ideal for laboratory sampling and testing needs. 1-888-777-7080


Issuu converts static files into: digital portfolios, online yearbooks, online catalogs, digital photo albums and more. Sign up and create your flipbook.