Vol 7 no 2 2013 by SURIC

ISSN 1905-9159

Silpakorn University

Science and Technology Journal Volume 7 Number 2 (July-December) 2013

Double Bootstrap-t One-sided ConďŹ dence Interval for Population Variance of Skewed Distributions Wararit Panichkitkosolkul

Diversity of Hyperthermophililic Bacteria Belonging to Order Thermotogales Thriving in Three Hot Springs in Thailand: Resources of Genes Encoding Thermostable Enzymes Porranee Keawram and Wirojne Kanoksilapatham

A Study of Adsorption of an Organic Colouring Matter on Powdered Natural Plant Material Kiran V. Mehta

Test Case Based Selection for the Process of Software Maintenance The subject program

Define a test suite

Generate the test case

Create the test case path The set of selected test cases

http://www.surdi.su.ac.th http://www.journal.su.ac.th

Select the test case

Adtha Lawanna

SILPAKORN UNIVERSITY Science and Technology Journal SUSTJ

Copyright All rights reserved. Apart from citations for the purposes of research, private study, or criticism and review, no part of this publication may be reproduced, stored or transmitted in any other form without prior written permission by the publisher. Published by Silpakorn University Printing House. Silpakorn University, Sanamchandra Palace Campus, Nakhon Pathom 73000 ÂŠ Silpakorn University ISSN 1905-9159

SILPAKORN UNIVERSITY Science and Technology Journal Editorial Office Silpakorn University Research and Development Institute (SURDI), Silpakorn University, Sanamchandra Palace Campus, Nakhon Pathom, Thailand

Editorial Policy All articles submitted for publication will be evaluated by a group of distinguished reviewers. The editorial board claims no responsibility for the contents or opinion expressed by the authors of individual article.

Editorial Advisory Board Assist. Prof. Alice Thienprasert, Ph.D Director, Silpakorn University Research and Development Institute, Thailand Prof. Amaret Bhumiratana, Ph.D Department of Biotechnology, Mahidol University, Thailand Prof. Geoffrey A. Cordell, Ph.D Professor Emeritus, University of Illinois at Chicago, USA Prof. Kanaya Shiginori, Ph.D Department of Material and Life Sciences, Osaka University, Japan Prof. Keiji Yamamoto, Ph.D Graduate School of Pharmaceutical Sciences, Chiba University, Japan Assoc. Prof. Kriengsak Poonsuk, Ph.D Faculty of Animal Sciences and Agricultural Technology, Silpakorn University, Thailand Assist. Prof. Lerkiat Vongsarnpigoon, Ph.D National Metal and Materials Technology Center (MTEC), Thailand Assoc. Prof. Nijsiri Ruangrungsri, Ph.D College of Public Health Sciences, Chulalongkorn University, Thailand Assoc. Prof. Petcharat Pongcharoensuk, Ph.D Department of Pharmacy, Mahidol University, Thailand Prof. Piyasan Praserthdam, Ph.D Department of Chemical Engineering, Chulalongkorn University, Thailand Assoc. Prof. Surachai Nimjirawath, Ph.D Department of Chemistry, Silpakorn University, Thailand Prof. Tharmmasak Sommartya, Ph.D Faculty of Agriculture, Bangkhen Campus, Kasetsart University, Thailand Prof. Virulh Sa-Yakanit, Ph.D Department of Physics, Silpakorn University, Thailand

Editor Assoc. Prof. Onoomar (Poobrasert) Toyama, Ph.D Faculty of Pharmacy, Silpakorn University

SILPAKORN UNIVERSITY Science and Technology Journal Editorial Board Assist. Prof. Bussarin Ksapabutr, Ph.D Faculty of Engineering and Industrial Technology, Silpakorn University Prof. Chawewan Ratanaprasert, Ph.D Faculty of Science, Silpakorn University Assist. Prof. Chockpisit Thepsithar, Ph.D Faculty of Science, Silpakorn University Assist. Prof. Jittat Fakcharoenphol, Ph.D Faculty of Engineering, Kasetsart University Assoc. Prof. Mana Kanjanamaneesathian, M.Appl.Sc. Faculty of Animal Sciences and Agricultural Technology, Silpakorn University Assist. Prof. Pramote Khuwijitjaru, Ph.D Faculty of Engineering and Industrial Technology, Silpakorn University Smith Tungkasmit, Ph.D College of Social Innovation, Rangsit University Assoc Prof. Soraya Ruamrungsri, Ph.D Faculty of Agriculture, Chiang Mai University Agnes Rimando, Ph.D U.S. Department of Agriculture, Agricultural Research Service, USA Prof. Juan Boo Liang, Ph.D Institute of Bioscience, Universiti Putra Malaysia, Malaysia Prof. Shuji Adachi, Ph.D Graduate School of Agriculture, Kyoto University, Japan Vincenzo Esposito, Ph.D Department of Energy Conversion and Storage, Technical University of Denmark, Denmark

Managing Editor Pranee Vichansvakul

Periodicity Twice yearly

All correspondence should be addressed to: Managing Editor, E-mail address: pranee_aon1@hotmail.com Telephone: 080-5996680

Information about the Journal An electronic journal is provided on the web sites (http://www.journal.su.ac.th, http://www.tci-thaijo.org/index.php/sustj/index, http://www.surdi.su.ac.th/journal_th.html). The journal is available at Silpakorn University Book Center. Telephone: 66-2223-7345, 66-2434-1792, 66-3424-4054.

Instructions to Authors Silpakorn University Science and Technology Journal (SUSTJ) is a peer review journal published twice a year in January and July by the Research and Development Institute of Silpakorn University, Thailand. SUSTJ puts together articles in Science and Technology and aims to promote and distribute peer reviewed articles in the areas of science, health science, animal science, agriculture, engineering, technology and related fields. Articles from local and foreign researchers, invited articles and review from experts are welcome.

Publication Ethics and Publication Malpractice Statement The Editor must ensure that each manuscript is initially evaluated by the editor/co-editor for the originality of the contents before forwarding to two referees for blind peer review. The review period will be from two weeks to one month (Maximum two months in extra ordinary circumstances). The Editorial Board evaluates manuscripts on their intellectual content without regard to authorsâ&#x20AC;&#x2122; race, gender, sexual orientation, or citizenship and must not disclose any information about a submitted manuscript to anyone other than the corresponding author, reviewers, potential reviewers, and the publisher.

Reviews should be conducted objectively without personal criticism.

Reviewers should express their views clearly with supporting arguments, identify relevant published work that has not been cited by the authors, and report any similarity or overlap between the manuscript under consideration and other published papers to the editorâ&#x20AC;&#x2122;s attention. Reviewers should not review manuscripts in which they have conflicts of interest resulting from competitive, collaborative, or other relationships or connections with any of the authors, companies, or institutions connected to the papers. The authors should ensure that they have written entirely original works, and has been appropriately cited or quoted the work of others. Manuscripts submitted must not have been published as copyrighted material elsewhere. Manuscripts under review must also not be submitted for consideration by other publication as copyrighted material. This should be declared in the Submission Form. By submitting a manuscript and a Submission Form, the author(s) agree that the copyright will be transferred to SUSTJ if the manuscript and associated multimedia are accepted for publication. The author(s) retain the rights to the fair use (e.g., teaching and non-profit uses) of the published material. Types of contributions Short communications, Research articles, Review articles

Preparation of manuscripts 1. The text should be double-spaced with line number on A4 and a font Times New Roman size 11 should be used. When using MS Word, insert all symbols by selecting “Insert-Symbol” from the menu and use the “Symbol” font. 2. Manuscripts should be organized in the following order: Cover page with title and authors’ names and affiliations Abstract (in English and Thai) Key Words Introduction Materials and Methods, Area Descriptions, Techniques Results Discussion Conclusion Acknowledgements References Tables and Figures Authors’ names and affiliations Full names and affiliations (marked with superscript number) should be provided for all authors on the cover page, separately from the content. The corresponding author (marked with superscript asterisk) should also provide a full postal address, telephone and fax number and an e-mail address as a footnote on the title page. Abstract First page of the content starts with Abstract, including title of the article on top of page. Provide a short abstract not more than 200 words, summarizing the question being addressed and the findings. Key Words Provide 3-5 key words or short phrases in alphabetical order, suitable for indexing. References In text references: Refer to the author’s name (without initials) and year of publication, e.g., Feldmann, 2004 (for 1 author), Feldmann and Langer, 2004 (for 2 authors), or Feldmann et al., 2004 (for more than 2 authors). Article references: References should be listed in alphabetical order of author(s). For journal, list all names of authors. Book Feldmann, H. (2004) Forty Years of FEBS, 2nd ed., Blackwell Publishing Ltd., Oxford, pp.1121-1129. Chapter in a book Langer, T. and Neupert, W. (1994) Chaperoning mitochondrial biogenesis. In The Biology of Heat Shock Proteins and Molecular Chaperones (Morimoto, R. I., Tissieres, A. and Georgopoulos, C., eds.), 3rd ed., pp. 53-83. Cold Spring Harbor Laboratory Press, Plainview, New York.

Article in a journal Hammerschlag, F. A., Bauchan, G., and Scorza, R. (1985) Regeneration of peach plants from callus derived from immature embryos. Journal of Natural Products 70(3): 248-251. Hammerschlag, F. A., Bauchan, G., and Scorza, R. Regeneration of peach plants from callus derived from immature embryos. Journal of Natural Products (in press). Article on the web Lee, K. (1999) Appraising adaptaive management. Conservation Ecology 3(2). [Online URL:www. consecolo.org/Journal/vol3/iss2/index.html] accessed on April 13, 2001. Proceedings MacKinnon, R. (2003) Modelling water uptake and soluble solids losses by puffed breakfast cereal immersed in water or milk. In Proceedings of the Seventh International Congress on Engineering and Food, Brighton, UK. Patent Yoshikawa, T. and Kawai, M. (2006) Security robot. U.S. Patent No. 2006079998. Tables and Figures Each Table and Figure must be on a separate page of the manuscript. Tables: Number the tables according to their sequence in the text. The text should include references to all tables. Vertical lines should not be used to separate columns. Leave some extra space instead. Figures: Figures should be of high quality (not less than 300 dpi JPEG or TIFF format), in black and white only, with the same size as the author would like them to appear in press. Choose the size of symbols and lettering so that the figures can be reduced to fit on a page or in a column. Submission of Manuscripts All information contained in a manuscript is a full responsibility of the authors, including the accuracy of the data and resulting conclusion. The editorial office will acknowledge receipt of the manuscript within 2 weeks of submission. The ‘accepted date’ that appears in the published article will be the date when the managing editor receives the fully revised version of the manuscript. The manuscript may be returned to authors for revision. Authors will be given 2 weeks after receipt of the reviewers’ comments to revise the manuscript. Please submit the manuscript with a submission form to the following address: e-mail: pranee_ aon1@hotmail.com Proofs Proofs will be sent to the corresponding author by e-mail (as PDF file) or regular mail. Author is requested to check the proofs and return any corrections within 2 weeks.

Silpakorn University Science and Technology Journal

Contents

Volume 7 Number 2 (July - December) 2013

Research Articles

Double Bootstrap-t One-Sided Confidence Interval for Population Variance

of Skewed Distributions...................................……....................……...................…….....................

Wararit Panichkitkosolkul

Diversity of Hyperthermophililic Bacteria Belonging to Order Thermotogales

Thriving in Three Hot Springs in Thailand: Resources of Genes

Encoding Thermostable Enzymes..........................................................................................…….... 17

Porranee Keawram and Wirojne Kanoksilapatham

A Study of Adsorption of an Organic Colouring Matter on

Powdered Natural Plant Material .......................................................................................……...... 29

Kiran V. Mehta

Test Case Based Selection for the Process of Software Maintenance ...............................……...... 36

Adtha Lawanna

Silpakorn University Science and Technology Journal (SUSTJ) is now available on the following databases: Chemical Abstract Service (CAS) SciFinder Scholar (CAPLUS) International Information System for the Agricultural Sciences and Technology (AGRIS) (FAO) AGRICultural Online Access (AGRICOLA) Food Science and Technology Abstracts (FSTA) The National Science Digital Library (NSDL) CAB Abstract Directory of Open Access Journals (DOAJ) Google Scholar Thai Journal Citation Index (TCI) Centre

Research Article Double Bootstrap-t One-Sided Confidence Interval for Population Variance of Skewed Distributions Wararit Panichkitkosolkul Department of Mathematics and Statistics, Faculty of Science and Technology, Thammasat University, Phathumthani, Thailand. Corresponding author. E-mail address: wararit@mathstat.sci.tu.ac.th Received February 4, 2013; Accepted May 7, 2013 Abstract This paper proposes a double bootstrap-t one-sided confidence interval for population variance of skewed distributions. The upper endpoint and lower endpoint confidence intervals are studied. The onesided confidence intervals based on the chi-square statistic, bootstrap-t method and double bootstrap-t method are compared via Monte Carlo simulations. The simulation results indicated that the coverage probabilities of bootstrap-t confidence interval can be increased by using double bootstrap resampling. The upper endpoint confidence interval using double bootstrap-t method predominates the other methods with respect to the coverage probability criteria. The performance of the proposed one-sided confidence interval is illustrated with an empirical example. Key Words: Double Bootstrap-t; Confidence Interval; Variance; Skewed Distribution Introduction A confidence interval (CI) for a population parameter gives a bound computed from sample data containing the true value of the parameter with a specified confidence level. Confidence interval plays a significant role in statistical inference regarding the parameter. For confidence interval for population variance, well-known existing methods are based upon the chi-square statistic which is introduced by Pearson (1900). Based on the chi-square statistic, the upper endpoint and lower endpoint (1 − α )100% confidence intervals 2 for σ are (Cojbasic and Loncar, 2011)

Silpakorn U Science & Tech J 7 (2) : 9-16, 2013

UCI χ 2

and

 (n − 1) S 2  = 0 , , χ n2−1,α  

 (n − 1) S 2  , + ∞ , LCI χ 2 =  2  χ n −1,1−α  n

2 (n − 1) −1 ∑ ( X i − X ) 2 , where S =

χ n2−1,1−α

are the

(α )100th

(2)

χ n2−1,α

i =1

and

(1)

and

(1 − α )100th

percentiles of the central chi-square distribution with n − 1 degrees of freedom. It is well-known that these upper endpoint and lower endpoint confidence intervals for σ 2 are constructed under the normal distribution. However, the underlying

ISSN 1905-9159

Silpakorn U Science & Tech J Vol.7(2), 2013

Double Bootstrap-t One-Sided Confidence Interval

given in the fifth section, and the conclusions are in distribution is non-normal in some situations. the sixth section. Hence, it may be a skewed distribution. To deal with these situations, many researchers have Bootstrap-t One-Sided Confidence Interval for proposed confidence interval for σ 2 of skewed distributions. For example, Bonett (2006) provided the Variance an approximate confidence interval for standard The bootstrap introduced by Efron deviation and his proposed confidence interval (1979) is a computer-based and resampling is nearly exact under the normal distribution method for assigning measures of accuracy for small samples, and under the non-normal to statistical estimates (Efron and Tibshirani, distributions for moderate samples. Cojbasic and 1993). For a sequence of independent and Tomovic (2007) presented confidence intervals identically distributed (i.i.d.) random variables, for the population variance and the difference the bootstrap procedure can be defined as follows in variances of two populations based on the (Tosasukul et al., 2009). Let X 1 , X 2 ,..., X n be ordinary t-statistics combined with the bootstrap independently and identically distributed random method. In addition, Cojbasic and Loncar (2011) variables from some distribution with mean studied the coverage accuracy of one-sided µ and variance σ 2 . Let the random variables bootstrap-t confidence intervals for the population { X *j ,1 ≤ j ≤ m} be the result from sampling m variances combined with Hall’s and Johnson’s times from the population with replacement from transformation. In some cases, we found that the the n observations X 1 , X 2 ,..., X n . The random above mentioned confidence intervals provided variables { X *j ,1 ≤ j ≤ m} are called the bootstrap the coverage probability less than the nominal samples from original data X 1 , X 2 ,..., X n . Let confidence interval. In this paper, we show n that a slight modification of the usual bootstrap S2 = (n − 1) −1 ∑ ( X i − X ) 2 be a sample variance. It i =1 confidence interval can help to improve the is well-known that the pivotal quantity (n − 1) S 2 / σ 2 accuracy of coverage probability. One approach of has central chi-square distribution with n − 1 increasing the coverage probability of confidence degrees of freedom (Bonett, 2006). A confidence interval is to use the double bootstrap described interval for population variance can be constructed by Nankervis (2002, 2005), which can be thought using aforementioned pivotal quantity. For large of as bootstrapping the bootstrap (Scherer and sample sizes, central chi-square distribution with Martin, 2005). Namely, the error in the coverage n − 1 degrees of freedom can be approximated by probability of bootstrap confidence intervals can be normal distribution with mean n − 1 and variance reduced by the use of double bootstrap confidence 2(n − 1) (Cojbasic and Tomovic, 2007). Therefore, intervals (Nankervis, 2002). the distribution of the standardized variable The structure of this paper is as follows. The next section presents the bootstrap-t one2 (n − 1) S − (n − 1) sided confidence interval for the variance, and 2 S2 −σ 2 σ Z = = (3) the third section provides the details of the double 2(n − 1) var( S 2 ) bootstrap-t one-sided confidence interval for the variance. The fourth section presents the Monte converges to standardize normal distribution as Carlo simulation results. An empirical example is n increases to infinity. One-sided bootstrap-t

W. Panichkitkosolkul

Silpakorn U Science & Tech J Vol.7(2), 2013

confidence interval for σ 2 is calculated based on Double Bootstrap-t One-Sided Confidence the statistic Interval for the Variance The details of double bootstrap-t one-sided S2 −σ 2 T= , (4) confidence interval are as follows. For each of  S2) var( B bootstrap replications, the first-level bootstrap samples { X *j ,1 ≤ j ≤ m} are first drawn from the original data. Next, the second-level bootstrap samples { X ** j ,1 ≤ j ≤ m} are drawn from the firstlevel bootstrap samples. The statistic T ** based on the second-level bootstrap samples is computed as follows

 S 2 ) is a consistent estimator of the where var( variance of S 2 . Casella and Berger (2001, pp.257) have shown the estimator of var( S 2 ) for nonnormal distribution such that  S 2 ) = 1  µˆ − n − 3 S 4  var(   n  4 n −1  = µˆ 4

and

T ** =

1 ∑ ( X i − X )4 . n i =1 n

where S **2

S **2 − S 2 ,  S **2 ) var(

(8)

is a standard deviation of the

{ X ** second-level bootstrap samples After re-sampling B bootstrap samples, in j }, each bootstrap sample we compute the value of the  S **2 ) = 1  µˆ ** − n − 3 S **4  var( and   following statistic n  4 n −1  ** 1 m S *2 − S 2 ( X i** − X ** ) 4 . Therefore, the upper = µˆ 4 * ∑ T = , (5) m i =1  S *2 ) var( endpoint and lower endpoint (1 − α )100% double

S *2

where statistic

bootstrap

replication

bootstrap-t confidence intervals for σ 2 are  S2) , UCI DB = 0 , S 2 + tˆ(1**−α ) var( (9)

 S *2 ) = 1  µˆ * − n − 3 S *4  and S 2 , var(   n  4 n −1 

(

1 m = µˆ 4* ∑ ( X i* − X * )4 . The upper endpoint and m i =1 lower endpoint (1 − α )100% bootstrap-t confidence intervals for σ 2 are

(

LCI DB = S + tˆ 2

** (α )

 var( S 2 ) , + ∞ , (10)

)

th ˆ** ˆ** where t(α ) and t(1−α ) are the (α )100 and  S2) , (6) var( (1 − α )100th percentiles of T ** given in Eq. (8).

)

UCI B = 0 , S 2 + tˆ(1* −α )

and

2 *  2 LCI B = S + tˆ(α ) var( S ) , + ∞ ,

(

)

where tˆ(*α ) and tˆ(1* −α ) are the (α )100th

(7)

Monte Carlo Simulation Results The following Monte Carlo experiment compares the performance of one-sided confidence intervals for the variance of skewed distributions. The simulation study was conducted using the open source statistical package R (Ihaka and Gentleman

and

(1 − α )100th percentiles of T * shown in Eq. (5).

Silpakorn U Science & Tech J Vol.7(2), 2013

Double Bootstrap-t One-Sided Confidence Interval

low (coefficients of skewness are equal to 0 and 0.62). Furthermore, the double bootstrap-t method provides the estimated coverage probabilities more than those of other methods. However, all methods have poor estimated coverage probabilities of upper endpoint confidence interval for medium and high skewness (coefficients of skewness are equal to 6.18 and 23.73). For instance, the estimated coverage probabilities of 95% upper endpoint confidence interval for Lognormal with Ď&#x192; 2 = 2 and n = 20 are 0.2872, 0.6047 and 0.7687

1996) to estimate the coverage probability of onesided confidence interval. We chose to use some of the probability density functions of Cojbasic and Loncar (2011) in the simulation study. For each probability density function, we generated ten thousand random samples from Weibull, Exponential and Lognormal distribution and used 2,000 bootstrap samples. The different sample sizes ( n = 10, 20, 50, 100) are considered. Table 1 illustrates the results of the estimated coverage probabilities of 95% lower endpoint confidence intervals while the estimated coverage probabilities of 95% upper endpoint confidence intervals are shown in Table 2. We begin with the results for the lower endpoint confidence intervals (Table 1). The chi-square method provides estimated coverage probabilities of the lower endpoint confidence intervals close to the nominal confidence level 0.95 when the coefficients of skewness are equal to 0 and 0.62. For example, when the underlying distribution is Weibull distribution with shape parameter 2, the estimated coverage probabilities of 95% lower endpoint confidence interval attained by the chisquare method are 0.9446, 0.9446, 0.9398 and 0.9349 for n = 10, 20, 50 and 100, respectively. In addition, the bootstrap-t method provides the estimated coverage probabilities close to the nominal confidence level 0.95 when the coefficients of skewness are equal to 0.62 and 2. The estimated coverage probabilities of lower endpoint confidence interval by using double bootstrap-t method are close to one as skewness coefficient gets larger. Next, the upper endpoint confidence intervals are considered (Table 2). The estimated coverage probabilities of upper endpoint confidence interval by using both chi-square and bootstrap-t method get reasonably close to the nominal confidence level 0.95 for low skewness is

by chi-square, bootstrap-t and double bootstrap-t methods, respectively. Additionally, all estimated coverage probabilities tend to increase as sample size gets larger. The above results indicate that the upper endpoint confidence interval using double bootstrap-t method dominates the other approaches for almost all situations except low skewness. An Empirical Example To illustrate an empirical example of onesided confidence intervals for population variance of skewed distributions that have been presented within the previous section, we have used the real environmental data. Sulfur dioxide (SO2) contents of air in micrograms per cubic meter for forty U.S. cities were collected from U.S. government publications. The data were obtained from 1969 to 1971 (Source: http://lib.stat.cmu.edu/DASL). The histogram, density plot, box plot and normal QQ plot of SO2 contents are displayed in Figure 1. It indicates that the distribution of SO2 contents was positively skewed. The 95% lower and upper endpoint confidence intervals for the variance are constructed. As shown in Table 3, the lower endpoint confidence intervals computed via double bootstrap-t method provides the widest length as compared to those obtained from chi-square and bootstrap-t method. It is corresponding with the Monte Carlo studies that the double

W. Panichkitkosolkul

Table 1

Silpakorn U Science & Tech J Vol.7(2), 2013

The estimated coverage probabilities of 95% lower endpoint confidence interval for the variance of standard normal distribution and skewed distributions. Distribution

Standard normal

Weibull with shape parameter 2

Exponential with mean 1

Lognormal with Ď&#x192; 2 = 1

Lognormal with Ď&#x192; 2 = 2

Method

Skewness coefficient

Sample size

Chi-square

Bootstrap-t

Double bootstrap-t

0.9506

0.8588

0.8965

0.9490

0.8942

0.9414

0.9511

0.9203

0.9687

100

0.9508

0.9328

0.9789

0.9446

0.9455

0.9755

0.9446

0.9541

0.9858

0.9398

0.9512

0.9899

100

0.9349

0.9440

0.9900

0.8805

0.9720

0.9877

0.8608

0.9777

0.9969

0.8416

0.9670

0.9975

100

0.8253

0.9610

0.9950

0.8918

0.9964

0.9991

0.8593

0.9953

0.9997

0.8200

0.9900

0.9998

100

0.7909

0.9852

0.9998

0.9221

0.9990

0.9994

0.8959

0.9991

1.0000

0.8642

0.9970

1.0000

100

0.8327

0.9945

0.9999

0.62

6.18

23.73

Silpakorn U Science & Tech J Vol.7(2), 2013

Double Bootstrap-t One-Sided Confidence Interval

Table 2 The estimated coverage probabilities of 95% upper endpoint confidence interval for the variance of standard normal distribution and skewed distributions. Distribution Standard normal

Weibull with parameter 2

shape

Exponential with mean 1

Lognormal with Ď&#x192; 2 = 1

Lognormal with Ď&#x192; 2 = 2

Method

Skewness coefficient

Sample size

Chi-square

Bootstrap-t

Double bootstrap-t

0.9460

0.9467

0.9940

0.9502

0.9503

0.9917

0.9476

0.9466

0.9910

100

0.9550

0.9561

0.9914

0.9492

0.9350

0.9911

0.9455

0.9274

0.9837

0.9438

0.9344

0.9829

100

0.9385

0.9404

0.9837

0.7981

0.8259

0.9398

0.7803

0.8505

0.9324

0.7793

0.8939

0.9577

100

0.7841

0.9118

0.9677

0.5340

0.7038

0.8432

0.4946

0.7233

0.8439

0.4973

0.7728

0.8711

100

0.5048

0.7977

0.8923

0.2960

0.5979

0.7410

0.2872

0.6047

0.7687

0.2888

0.6395

0.7792

100

0.3029

0.6727

0.7928

0.62

6.18

23.73

Conclusions A double bootstrap-t one-sided confidence interval for population variance of skewed distributions has proposed in this paper. The study was carried out to compare the performance of a proposed confidence interval with the existing confidence intervals. Three one-sided confidence intervals are considered: the one-sided confidence interval based on chi-square statistic, the bootstrap-t

one-sided confidence interval and the double bootstrap-t one-sided confidence interval. Based on simulation studies, the double bootstrap-t onesided confidence interval provides good coverage probability for the upper endpoint confidence interval. On the other hand, the double bootstrap resampling can also improve the accuracy of the upper endpoint confidence interval for population variance of skewed distributions. The behind

W. Panichkitkosolkul

Silpakorn U Science & Tech J Vol.7(2), 2013

bootstrap-t method provides the estimated coverage probabilities more than those of other methods. Therefore, the double bootstrap-t method is not suitable for this case. However, the length of Table 3

the upper endpoint confidence interval computed by double bootstrap-t method is shorter than other confidence intervals.

The 95% lower and upper endpoint confidence intervals for the variance of SO2 contents. Method

Lower endpoint confidence interval

Upper endpoint confidence interval

Chi-square

[ 0 , 831.33 ]

[ 395.24 , ∞ ]

Bootstrap-t

[ 0 , 1396.97 ]

[ 298.09 , ∞ ]

Double bootstrap-t

[ 0 , 1802.90 ]

[ 204.10 , ∞ ]

Figure 1 (a) Histogram (b) Density plot (c) Box plot and (d) Normal QQ plot of Sulfur dioxide(SO2) contents of air for forty-one US cities.

Silpakorn U Science & Tech J Vol.7(2), 2013

Double Bootstrap-t One-Sided Confidence Interval

Efron, B. (1979) Bootstrap methods: Another look at the jackknife. Annals of Statistics 7(1): 1-26. Efron, B. and Tibshirani, R. J. (1993) An Introduction to the Bootstrap. Chapman & Hall, New York. Hall, P. (1986) On the bootstrap and confidence intervals. Annals of Statistics 14(4): 14311452. Ihaka, R. and Gentleman, R. (1996) “R: A Language for Data Analysis and Graphics.” Journal of Computational and Graphical Statistics 5: 299-314. Nankervis, J. C. (2002) Stopping rules for double bootstrap confidence intervals, [Online URL: www.citeseerx.ist.psu.edu/viewdoc/ download?doi=10.1.1.86.8480&rep=rep1& type=pdf] accessed on November 19, 2012. Nankervis, J. C. (2005) Computational algorithms for double bootstrap confidence intervals. Computational Statistics & Data Analysis 49(2): 461-475. Pearson, K. (1900) On the criterion that a given system of deviations from the probable in the case of a correlated system of variables is such that it can be reasonably supported to have arisen from random sampling. Philosophical Magazine 50(5): 157-175. Scherer, B. and Martin, R. D. (2005) Introduction to Modern Portfolio Optimization with NUOPT and S-PLUS. Springer, New York. Tosasukul, J., Budsaba, K., and Volodin, A. (2009) Dependent bootstrap confidence intervals for a population mean. Thailand Statistician 7(1): 43-51.

reason is that the resulting double bootstrap confidence intervals have been shown to have a smaller order of error. For example, Hall (1986) has shown that, in general, the coverage rate of a 100(1-2α)% equal-tailed bootstrap confidence interval is corrected from 1-2α + O(n-1) to 1-2α + O(n-2) for a double bootstrap confidence interval. In addition, the coverage probability of double bootstrap-t lower endpoint confidence interval does not achieve exactly the nominal confidence level. Acknowledgements The authors thank the anonymous referees for their constructive suggestions and comments that resulted in an improved present of this paper. References Bonett, D. G. (2006) Approximate confidence interval for standard deviation of nonnormal distributions. Computational Statistics & Data Analysis 50(3): 775-782. Casella, G. and Berger, R. L. (2001) Statistical Inference. Duxbury Press, Pacific Grove, pp.257. Cojbasic, V. and Loncar, D. (2011) One-sided confidence intervals for population variances of skewed distribution. Journal of Statistical Planning and Inference 141(5): 1667-1672. Cojbasic, V. and Tomovic, A. (2007) Nonparametric confidence intervals for population variances of one sample and the difference of variances of two samples. Computational Statistics & Data Analysis 51(12): 5562-5578.

Research Article Diversity of Hyperthermophilic Bacteria Belonging to Order Thermotogales Thriving in Three Hot Springs in Thailand: Resources of Genes Encoding Thermostable Enzymes Porranee Keawram and Wirojne Kanoksilapatham* Department of Microbiology, Faculty of Science, Silpakorn University, Nakhon Pathom, Thailand * Corresponding author. E-mail: wirojne@su.ac.th Received January 21, 2013; Accepted June 6, 2013 Abstract Hyperthermophilic microorganisms belonging to order Thermotogales and thriving in high temperature habitats grow at temperatures above 70 °C to near 100 °C. In this study, 15 rod shaped bacteria with characteristic of terminal sac-like membrane were isolated from Pong Duet, Fang, and Jae Son hot springs. Differential biochemical properties of the isolates were characterized. Twelve isolates were detected utilizing carbohydrates in CT basal medium, but no growth was observed in the other three isolates. Keratin in native duck feather was degraded at 75-80 °C by 8 isolates. Based on morphology, physiology, and 16S rDNA sequences, 5 and 10 isolates belonging to morphological groups I and II were identified as Thermotoga species and Fervidobacterium species, respectively. The 16S rDNA type II PCR profile of isolates FC2004, FC201, FC202, FA004 and JS602 distinguish themselves from previously reported known species belonging to Fervidobacterium. Results obtained from this study indicate that some of Thailand’s isolates are distinct, and the geothermal spring ecosystems are rich in divergent hyperthermophiles which still remain to be explored. The hyperthermophilic isolates are crucial sources of numerous thermostable enzymes with potential to be applicable in the degradation of polymers in agricultural wastes such as starch, cellulose, and keratin. Key Words: Hot spring; Hyperthermophile; Thermotogales; Thermostable Enzyme Introduction Hyperthermophilic archaea and bacteria have been recognized as the most primitive forms of life on earth. Among the bacterial members, the thermophiles and hyperthermophiles belonging to order Thermotogales were phylogenetically positioned in close proximity to those ancient Archea (Stetter, 1996; Woese et al., 1990; Huber et al., 1986; Fiala and Stetter, 1986; Patel et al., 1985). Typical ecosystems suitable for growth

Silpakorn U Science & Tech J 7 (2) : 17-28, 2013

of hyperthermophilic bacteria belonging to order Thermotogales are extremely hot environments such as marine hydrothermal systems, petroleum reservoirs, and continental hot springs. Members of this order are obligately anaerobic heterotrophs growing on various complex substrates such as proteins, starch, cellulose, and xylan. Many of them have been demonstrated gaining energy from anaerobic respiration using elemental sulfur (S°) and sulfur compounds (Miranda-Tello et al., 2004; Balk

ISSN 1905-9159

Silpakorn U Science & Tech J Vol.7(2), 2013

Diversity of Hyperthermophilic Bacteria

islandicum (Huber et al., 1990) and F. gondwanense (Andrews and Patel, 1996) were isolated from a volcanic hot spring and geothermal artesian water, respectively. Both F. islandicum (opt. temp. of 65 °C) and F. gondwanense (opt. temp. of 65-68 °C) failed to grow at ≤45 °C; however, a maximum growth temperature close to 80 oC was reported. All isolates were reported growing in media with dilute NaCl concentration (opt. conc. of 0.1-0.2 % NaCl), and not growing at concentrations of > 0.6 - 1% NaCl. In addition, two hyperthermophiles growing optimally at >70 °C named Fervidobacterium pennivorans (opt. temp. of 70 °C) and F. changbaicum (opt. temp. of 80 °C) were previously isolated from volcanic hot springs. Both species required diluted NaCl concentration (0-0.1 % NaCl) for optimal growths (Cai et al., 2007; Friedrich and Antranikian, 1996). In this study, hyperthermophilic microorganisms were successfully isolated from three hot springs located at different angular distances in Northern Thailand. The isolates are sources of genes encoding numerous thermostable enzymes. The attempt to culture microorganisms growing at a temperature near the boiling point of water is considered groundbreaking in Thailand.

et al., 2002; Wery et al., 2001; Ravot et al., 1995; Patel et al., 1985). Currently, order Thermotogales comprises single family Thermotogaceae, which consists of eleven official genera. Typically, all members belonging to this order are gram-negative nonendospore forming rods. Cells are usually surrounded by a thick membranous sheath and form a sac-like structure (or a toga) at one or both terminals (Jayasinghearachchi and Lal, 2011; Feng et al., 2010; L’Haridon et al., 2002; Davey et al., 1993; Huber et al., 1989). Most members belonging to order Thermotogales were reported as moderate thermophiles growing at temperature around 60 °C or below (Miranda-Tello et al., 2004; Jayasinghearachchi and Lal, 2011; Alain et al., 2002; Wery et al., 2001; Jeanthon et al., 1995). Few genera have been reported growing in the temperature range of mesophiles (Nesbo et al., 2012; Dipippo et al., 2009). Some species belonging to the genera Thermotoga spp. and Fervidobacterium spp. have been frequently reported growing at the temperature of hyperthermophiles (70 °C to 90 °C). Thermotoga maritima and T. neapolitana were first isolated from marine hydrothermal ecosystems and later were found in low salt continental hot springs. Both species were recorded to grow at the upper growth limit temperatures up to 90 °C (Huber et al., 1986). Both Thermotoga petrophila and T. naphthophila were discovered from a subterranean oil reservoir in Japan and grew optimally at 80 °C. Both of them tolerated high salt concentration of > 5% (w/v) NaCl and required at least 0.1% NaCl for growths (opt. conc. of 1% NaCl) (Takahata et al., 2001). Fervidobacterium nodosum, the representative new genus, was first isolated from volcanic hot springs in New Zealand, and its growth temperature range of 47 to 80 °C (opt. temp. of 70 °C) was reported (Patel et al., 1985). Fervidobacterium

Materials and Methods Sample Collection and Isolation of Hyperthermophiles Sediment samples were collected near thermal sources of Pong Duet (19° 7´ N, 98° 56´ E), Fang (19° 58´ N, 99° 12’ E), and Jae Son (18° 50´ N, 99° 28´ E) hot springs in Northern Thailand. The temperatures measured in situ at the thermal exits were 80 to 100 °C. The sediments were transported to a laboratory in an ice box. Approx. 1 g of samples were inoculated into 480G medium in serum bottles and incubated anaerobically at 75 to 80 °C for 1 to 2 days or until turbidity was observed. Isolation of pure cultures was performed using serial tube

P. Keawram and W. Kanoksilapatham

Silpakorn U Science & Tech J Vol.7(2), 2013

Ability to Degrade Duck Feather A medium named FD was modified from medium I described by Friedrich and Antranikian (1996). The medium was employed to test the feather degradation capacity. A liter of the FD medium was composed of K2HPO4.3H2O (2.09

dilution technique for at least triplicate times. Pure cultures obtained were named after the initial letter of the hot springs’ names. Briefly, FA0, FC0, FC1, FC2 and FC3 indicate the wells located at Fang hot spring. PD5 stands for well 5 located at Pong Duet hot spring. JS4, JS5 and JS6 stand for wells 4, 5 and 6 located at Jae Son hot spring. The last two or three digits indicate isolate numbers. All cultures were stored in 480G medium at 4 °C. 480G Medium and Preparation A liter of 480G medium was composed of NaCl (0.5 g), NH4Cl (0.33 g), CaCl2.2H2O (0.15 g), MgCl2.6H2O (0.35 g), KCl (0.3 g), KH2PO4 (0.3 g), pancreatic digestion of casein (1 g), yeast extract (0.5 g), A5 solution (1 ml), resazurin solution (0.5 ml of 0.2 g/l solution) and Na2S.9H2O solution [3 ml of 25% (w/v), pH 7]. pH was adjusted to 7.27.5 at room temperature using 1N NaCl or 1N HCl before sterilization. The medium was prepared anaerobically in serum bottles under N2 atmosphere. Sterilization was performed at 100 °C for 1 h. A liter of the A5 solution was composed of Co(NO3)2.6H2O (0.00494 g), CuSO4.5 H2O (0.0079 g), H3BO3 (0.286 g), MnCl2.4 H2O (0.181 g), Na2MoO4.2H2O (0.039 g) and ZnSO4.7H2O (0.0222 g). Carbohydrate Utilization Test CT medium, a basal medium for testing carbohydrate utilization, was developed in this study. Compositions of the medium were similar to those of the 480G medium except that 0.1 g/l of pancreatic digestion of casein and 0.05 g/l of yeast extract were employed. Glucose, sucrose, maltose, lactose, starch, cellobiose and carboxymethyl cellulose (CMC) were tested at a final concentration of 1 g/l. Briefly, overnight inoculum was diluted 100 times using the basal medium prior to the inoculation (to obtain an approx. conc. of 105cells/ml) into the triplicate bottles of a test sugar. Cell yields compared with those of the controls were determined at 48 h. Cells were counted using the direct count technique or measured optical density (OD660nm).

g), NaH2PO4.2H2O (1.29 g), (NH4)2SO4 (1.5 g), CaCl2.2H2O (0.1 g), NaCl (0.3 g), MgSO4.7H2O (0.3 g), NaHCO3 (1 g), pancreatic digestion of casein (1 g), yeast extract (1 g), cysteine hydrochloride (0.5 g), FeCl2.6H2O (6 mg), A5 solution (1 ml), resazurin solution (0.5 ml of 0.2 g/l solution) and Na2S.9H2O solution [3 ml of 25% (w/v), pH 7]. pH was adjusted to 7.2-7.5 at room temperature using 1N NaCl or 1N HCl before sterilization. The medium was prepared in a Hungate tube (15 ml) that contained a piece of duck feather (15 mg). Cultures were incubated at 75-80 °C for 48 h. Primers and PCR Conditions 16S DNA fragments were amplified using the following primers: THER3F, A109F, U515F, 940EcoRIrc and UA1406R. The THER3F primer was designed in this study and has the nucleotide sequence of 5’ AGGGTTTGATCMTGG 3’. Nucleotide sequence of 940EcoRIrc (5’ CGGCGTGAATTCCAATTAAACCGCACGC 3’) was previously described (Kanoksilapatham et al., 2012). Nucleotide sequences of A109F (5’ ACKGCTCAGTAACACGT 3’), U515F (5’ GTGCCAGCMGCCGCGGTAA 3’) and UA1406R (5’ ACGGGCGGTGWGTRCAA 3’) were described elsewhere (Baker and Cowan, 2004). Relative binding positions of these primers on a 16S rDNA sequence (GenBank AE000512) are shown in Figure 1. Predicted from the binding sites of the 940EcoR1rc primer, either a 1000 bp-long PCR product or none (Figure 1a) might be amplified. The A109F, an archaeal specific primer, is anticipated to bind more strongly on the 16S rDNA sequences from Thermotoga than Fervidobacterium (Figure 1b).

Silpakorn U Science & Tech J Vol.7(2), 2013

Diversity of Hyperthermophilic Bacteria

and PD522), 9 isolates from Fang (FA002, FA003, FA004, FC1002, FC2004, FC201, FC202, FC203 and FC303), and 3 isolates from Jae Son (JS401, JS504 and JS602) hot springs. Phase contrast micrographs reveal that the cells of these isolates share common characteristics of rod shaped with a terminal toga at one or both ends (Figure 2), enclosed in membranous sheath, stained with Gram-negative or Gram-nonreactive, usually arranged singly and rarely in pairs to short chain. Occasionally filamentous cells were detected on some isolates (Figures 2c-2d). No endospore was observed on all isolates. Based on characteristic of toga and shape,

All PCR reactions were performed as follows: 1 cycle of 95°C for 4 min then 30 cycles of 95°C for 1 min, 57°C for 1 min and 72°C for 1 min; and a final extension at 72°C for 10 min. The products were separated using 1% agarose gel electrophoresis. Results and Discussion Isolation and Morphology Fifteen strictly anaerobic heterotrophs growing at temperature of 80 °C were isolated from three hot springs (Pong Duet, Fang, and Jae Son hot springs) located in Northern Thailand. They include 3 isolates from Pong Duet (PD501, PD502

Figure 1 Relative binding positions of primers on 16S rDNA sequences (1.5 kb). (a) Diagram shows relative binding positions of primers on the 16S rRNA gene of T. maritima (GenBank AE000512). Nt. no. 1 corresponds to the nt. no. 188970 of the AE000512. GenBank JF339224 which was amplified by a 940EcoR1rc is shown. An inverted sequence of 940EcoR1rc (named 940EcoR1rc?) located at upstream of the JF339224 is identified in this study. Binding site of THER3 primer begins at nt. no.23 of the JF339224 and the nt. no. 6 of the AE000512. Relative binding sites of universal primers (U515F, 940EcoR1rc and UA1406R) are indicated at relative positions. (b) Alignment of A109F primer sequence (5’ ACKGCTCAGTAACACGT 3’) versus sequences from order Thermotogales was conducted using MEGA5.5. Symbols: Horizontal boxes indicate 16S rRNA genes. Arrows represent primers and arrow heads indicate direction of PCR polymerization. Underline indicates identical bases.

P. Keawram and W. Kanoksilapatham

Silpakorn U Science & Tech J Vol.7(2), 2013

Biochemical Properties Ability to utilize carbohydrate of the isolates was tested in CT basal medium containing 1 g/l of test carbohydrates. In general, slight growths (with a magnitude of 106 cells/ml) were observed in the CT basal medium (controls), but no growth was detected on the isolates FA003, PD522 and FC202 (data not shown). In addition, no growth on all carbohydrates tested was also observed in these three isolates, implying this medium might not be suitable for their growths (Table 1). Therefore, these isolates are categorized into a separate group named in this study as “biochemical subgroup I”, and the remaining 12 isolates are grouped as “biochemical subgroup II”. Members of the subgroup II were

they could be categorized into two morphological groups. The first group includes the isolates FA002, FA003, FC203, FC1002 and PD522. Cells of the first group are long slender rods (an average size range of 0.4-0.5x1.5-5 μm) with a toga at both terminals; this characteristic is a typical hallmark of the genus Thermotoga spp. (Figures 2a and 3a). The second group includes the isolates PD501, PD502, FA004, FC2004, FC201 FC203, FC303, JS401, JS504 and JS602. Cells of the second group are short rods (size range of 0.5-0.6x1-2.5 μm) with a balloon-like toga presenting at a terminal (Figures 2b-2d and 3b-3c). Few long filaments (an average size range of 10-40 μm) were rarely observed on some isolates (Figures 2c-2d).

Figure 2 Phase contrast micrographs of some isolates belonging to order Thermotogales. (a) Isolate FA002 shows slender rod shaped cells with a toga at both terminals. Cells occur singly sized of 0.4-0.5x2.5-5 μm. Arrow heads indicate toga. (b) Isolate FC2004 shows short rod shaped cells with a single large terminal toga. Cells are usually encased in thick sheath-like membrane, occurring singly or in pairs. Arrow heads indicate toga. (c) Isolate JS401 shows short rod shaped cells with a toga at one terminal. Filament with a terminal spheroid toga is occationally detected. Cells arrange singly and short chain. Arrow head indicates a filamentous cell with a balloon-like toga. (d) Isolate FC2004 shows a long filamentous and short rod shaped cells with a terminal toga. Arrows indicate toga.

Silpakorn U Science & Tech J Vol.7(2), 2013

Diversity of Hyperthermophilic Bacteria

Fervidobacterium changbaicum (Cai et al., 2007), and Fervidobacterium islandicum (Nam et al., 2002) digested keratin in feather. Ability to degrade native feather of the 15 isolates was tested at 75 °C and 80 °C. Results reveal that all of the 5 isolates belonging to the morphological group I were found unable to degrade the feather. In contrast, 8 from 10 isolates belonging to the morphological group II significantly degraded keratin in the feather (Table 1). Although isolate FC202 was inert on the test carbohydrates, it was observed degrading the feather in the medium.

detectable growing at least on glucose, sucrose and maltose (with a magnitude of 107 cells/ml). No growth on lactose was detected in all isolates, except that slight growth was observed on isolate FC1002. Little to remarkable growths on soluble starch and carboxymethyl cellulose (CMC) were demonstrated on isolates FC1002, FC2004, FA004, FC201, JS602 and FC303 (Table 1). Degradation of Duck Feather Previous reports suggested that some strains of hyperthermophilic Fervidobacterium pennavorans (Friedrich and Antranikian, 1996),

Table 1 Biochemical properties of hyperthermophilic isolates belonging to order Thermotogales. All tests were performed in triplicates and compared with controls.

Sucrose

Maltose

Lactose

Cellobiose

Soluble starch

CMC

I II I II II II I II II II II II II

III III III III III

Biochemical

numbers

groups

subgroup

FA002

FA003

FC203

PD522

FC2004

FA004

FC201

FC202

JS602

FC303

PD501

PD502

JS401

JS504

native feather2

Glucose +

Morphological

Types

II II

Isolates

FC1002

Degradation of

Carbohydrate utilization1

+ = growth, ± = slight growth, - = no growth, (n=3).

+ = degrade duck feather within 48 h, - = not degrade duck feather within 48 h, (n=3).

of 16S rDNA profiles

P. Keawram and W. Kanoksilapatham

Silpakorn U Science & Tech J Vol.7(2), 2013

16S rDNA Profiles 16S rDNA profiles were constructed using three primer pairs (U515F/UA1406R, A109F/940EcoR1rc and THER3F/UA1406R). Approx. sizes of 900 and 1400 bp-long PCR products were anticipated from the binding positions of the U515F/UA1406R and THER3F/UA1406R, respectively (Figures 1a). Analysis of the A109F priming sequences (length at 3’end) suggests that the primer might bind more strongly to the sequences of Thermotoga spp. than Fervidobacterium spp. (Figure 1b). Experimental results reveal three distinguishable profiles generated using these primer pairs (Figure 4). As expected, the bands with approx. size of 900 and 1400 bp-long were obtained from all isolates when amplified using the universal primer pairs of U515F/UA1406R and THER3F/UA1406R, respectively. On the other hand, 3 distinct PCR profiles (Figure 4) named “16S type I”, “16S type II”, and “16S type III” were revealed using the

A109F/940EcoR1rc. The 16S type I is identified by a faint PCR product size of 850 bp-long amplified using the A109F/940EcoR1rc (lane 3 in Figure 4), and it was disclosed on the isolates belonging to genus Thermotoga spp. (FC1002, FA002, FA003, FC203 and PD522) mentioned above (see also alignment of the A109F in Figure 1b). The 16S rDNA sequence of isolate FC1002 (GenBank JF339227) reveals the highest similarity (94-97%) to several known sequences of Thermotoga spp. (GenBank nos. AE000512, AJ401024, CP001839, CP000702 and NR_024751). The results confirm that the isolates with a toga at both terminals share the same characteristic as genus Thermotoga species (Figure 3a). The 16S type II is depicted by a prominent DNA fragment size of 1000 bp-long amplified using the A109F/940EcoR1rc (lane 6 in Figure 4), and it was demonstrated on 5 isolates belonging to genus Fervidobacterium (FC2004, FA004, FC201,

Figure 3 Classification based on morphology and 16S profiles. (a) Cells with morphological group I and 16S type I profile are classified as genus Thermotoga species. Isolate numbers were listed beneath the pictures. (b) Cells with morphological group II and 16S type II profile are classified as genus Fervidobacterium species. Isolate numbers were listed beneath the pictures. (C) Cells with morphological group II and 16S type III profile are classified as genus Fervidobacterium species. Isolate numbers were listed beneath the pictures.

Silpakorn U Science & Tech J Vol.7(2), 2013

Diversity of Hyperthermophilic Bacteria

FC202 and JS602). The 16S rDNA sequence of the isolate FC2004 (GenBank JF339226) reveals 96% similarity to Fervidobacterium changbaicum strain CBS-1 (GenBank EF138832), F. islandicum strain AW-1 (GenBank AF434670) and F. nodosum Rt17-B1 (GenBank CP000771) and 92% similarity to F. gondwanense strain AB39 (GenBank NR_036997). The results confirm these isolates as Fervidobacterium spp. (Figure 3b). In addition, a reverse priming site of the 940EcoR1rc (named 940EcoR1rc?) is identified in this study (Figures 1a). In order to confirm the presence of this priming site, PCR amplification reactions using the single 940EcoR1rc primer and DNA templates from the 16S type II isolates reveals 1000 bps PCR product (data not shown). However, this additional priming site (940EcoR1rc?), at the adjacent sequences of the 16S rRNA genes, was absent in the complete genome sequences of F. nodosum (GenBank CP000771) and F. penivorans (GenBank CP003260). The results imply that these 5 isolates are differentiated from F. nodosum and F. penivorans.

The 16S type III is recognized by lacking PCR product when amplified using the A109F/940EcoR1rc (lane 9 in Figure 4), and it was determined on the other 5 isolates belonging to genus Fervidobacterium (FC303, PD501, PD502, JS401 and JS504). The missing DNA band might result from a weak bonding (where the 3â&#x20AC;&#x2122; hydroxyl end of the A109F primer) on several reported 16S rDNA sequences from Fervidobacterium species including the F. nodosum and F. penivorans mentioned above (Figure 1b). Diversity of the Hyperthermophilic Bacteria across the Hot Springs Five isolates belonging to Thermotoga spp. were discovered from Fang and Pong Duet hot springs (four and one isolates), respectively (Figure 3a). However, no Thermotoga sp. was obtained from Jae Son hot spring. Unlike the isolate FC1002, the isolate FA002 is unable to utilize lactose, soluble starch and CMC, and the isolate FC203 is unable to utilize lactose and CMC (Table 1). Although, the isolates FA003 and PD522, which are classified

Figure 4 16S rDNA profiles generated using 3 primer pairs (U515F/UA1406R, A109F/940EcoR1rc and THER3F/UA1406R). Lane 1 indicates 100 bp ladder size markers. Lanes 2-4 represent profile type I that was amplified from the isolates FC203, FC1002, FA002, FA003 and PD522. Lanes 5-7 represent a profile type II that was amplified from the isolates FC2004, FA004, FC201, FC202 and JS602. Lanes 8-10 represent profile type III that was amplified from the isolates FC303, PD501, PD502, JS401 and JS504.

P. Keawram and W. Kanoksilapatham

Silpakorn U Science & Tech J Vol.7(2), 2013

utilizing carbohydrates including soluble starch and CMC. Five isolates (FC1002, FA002, FA003, FC203 and PD522) were identified as the genus Thermotoga species (Figure 3a), and ten isolates as the genus Fervidobacterium species (Figures 3b and 3c). It is noticed that the isolates belonging to the genus Thermotoga species do not degrade native feather. In contrast, 8 of the 10 isolates belonging to the genus Fervidobacterium species substantially degraded keratin excluding the isolates FC201 and FC303 (Table 1). In conclusion, geothermal hot spring ecosystems in Thailand are rich in cultured hyperthermophilic species belonging to Thermotoga and Fervidobacterium. Strain differentiation among

within the biochemical subgroup I, are inert on carbohydrate utilization, they were isolated from remote habitats. The results suggest that all 5 isolates belonging to Thermotoga spp. are distinct in their biochemical characteristics. Ten isolates with the morphological group II are identified as Fervidobacterium spp. (Figures 3b and 3c). They include 5 isolates from Fang (FC2004, FA004, FC201, FC202 and FC303), 2 isolates from Pong Duet (PD501 and PD502), and 3 isolates from Jae Son hot springs (JS401, JS504 and JS602). Excluding the isolates FC201 and FC303, all Fervidobacterium spp. were keratindegrading detectable in native feather. Unlike isolate FC2004, isolates FA004, FC201 and JS602 are unable to utilize cellobiose. In contrast to isolate PD501, isolate PD502 is unable to utilize cellobiose. In contrast to isolate JS401, isolate JS504 utilizes maltose, but not cellobiose. Among isolates belonging to Fervidobacterium spp., isolate FC202 was determined inert on carbohydrate utilization (Table 1). However, it degrades keratin in native feather. The 16S type II and III profiles imply diverse sequences of these isolates belonging to Fervidobacterium. In this study, the 16S type II profile is uniquely determined only in some Thai strains.

these isolates using conserved sequences from tRNA genes and an arbitrarily primed PCR based technique is under investigation (Welsh and McClelland, 1991; Patlada et al., 2011). Among the three hot springs examined, Fang hot spring might be a unique habitat and suitable for growths of divergent hyperthermophiles. The 16S type II observed among the 50% of strains belonging to Fervidobacterium indicates that they are differentiated from those reported known Fervidobacterium nodosum and F. pennivorans (Friedrich and Antranikian, 1996; Patel, et. al., 1985). The hyperthermophilic isolates obtained from this study are crucial sources of thermostable enzymes with potential to be applicable in degrading polymers such as starch, cellulose and insoluble keratin.

Conclusions Tw o m o r p h o l o g i c a l g r o u p s o f hyperthermophilic bacteria belonging to order Thermotogales were isolated from various sediment samples collected from Pong Duet (3 isolates), Fang (9 isolates), and Jae Son (3 isolates) hot springs. All grew at temperatures around 80 oC by gaining carbon and energy from pancreatic digest of casein and yeast extract. Strain differentiation was revealed based on biochemical properties and 16S rDNA fragment profiles. Twelve isolates were determined,

Acknowledgements This work was financially supported by grants from the Scientific Promotion and Development Fund, Faculty of Science, Silpakorn University (RGI 2553-06) and Silpakorn University Research and Development Institute (SURDI 54/01/18 and SURDI 55/01/05).

Silpakorn U Science & Tech J Vol.7(2), 2013

Diversity of Hyperthermophilic Bacteria

References Alain, K., Marteinsson, V. T., Miroshnichenko, M. L., Bonch-Osmolovskaya, E. A., Prieur, D., and Birrien, J. -L. (2002) Marinitoga piezophila sp. nov., a rod-shaped, thermopiezophilic bacterium isolated under high hydrostatic pressure from a deep-sea hydrothermal vent. International Journal of Systematic and Evolutionary Microbiology 52: 1331-1339. Andrews, K. T. and Patel, B. K. C. (1996) Fervidobacterium gondwanense sp. nov., a new thermophilic anaerobic bacterium isolated from nonvolcanically heated geothermal waters of the Great Artesian Basin of Australia. International Journal of Systematic Bacteriology 46(1): 265-269. Baker, G. C. and Cowan, D. A. (2004) 16S rDNA primers and the unbiased assessment of thermophile diversity. Biochemical Society Transactions 32(2): 218-221. Balk, M., Weijma, J., and Stams, A. J. (2002) Thermotoga lettingae sp. nov., a novel thermophilic, methanol-degrading bacterium isolated from a thermophilic anaerobic reactor. International Journal of Systematic and Evolutionary Microbiology 52: 13611368. Cai, J., Wang, Y., Liu, D., Zeng, Y., and Xue, Y. (2007) Fervidobacterium changbaicum sp. nov., a novel thermophilic anaerobic bacterium isolated from a hot spring of the Changbai Mountains, China. International Journal of Systematic and Evolutionary Microbiology 57(Pt 10): 2333-2336. Davey, M. E., Wood, W.A., Key, R., Nakamura, K. and Stahl, D. (1993) Isolation of three species of Geotoga and Petrotoga: two new genera, representing a new lineage in the bacterial line of descent distantly related to

the “Thermotogales”. Systematic and Applied Microbiology 16: 191-200. Dipippo, J. L., Nesbo, C. L., Dahle, H., Doolittle, W. F., Birkland, N. K., and Noll, K. M. (2009) Kosmotoga olearia gen. nov., sp. nov., a thermophilic, anaerobic heterotroph isolated from an oil production fluid. International Journal of Systematic and Evolutionary Microbiology 59: 2991-3000. Feng, Y., Cheng, L., Zhang, X., Li, X., Deng, Y., and Zhang, H. (2010) Thermococcoides shengliensis gen. nov representing a novel genus of the order Thermotogales from oilproduction fluid. International Journal of Systematic and Evolutionary Microbiology 60: 932-937. Fiala, G. and Stetter, K. O. (1986) Pyrococcus furiosus sp. nov. represents a novel genus of marine heterotrophic archaebacteria growing optimally at 100 °C. Archives of Microbiology 145: 56-61. Friedrich, A. B. and Antranikian, G. (1996) Keratin Degradation by Fervidobacterium pennavorans, a Novel Thermophilic Anaerobic Species of the Order Thermotogales. Applied and Environmental Microbiology 62(8): 2875-2882. Huber, R., Langworthy, T. A., Kölnig, H., Thomm, M., Woese, C. R., Sleytr, U. B., and Stetter, K. O. (1986). Thermotoga maritima sp. nov. represents a new genus of uniquely extremely thermophilic eubacteria growing up to 90°C. Archives of Microbiology 144: 324-333. Huber, R., Woese, C. R., Langworthy, T. A., Fricke, H., and Stetter, K. O. (1989) Thermosipho africanus gen. nov. represents a new genus of thermophilic eubacteria within the ‘‘Thermotogales’’. Systematic and Applied Microbiology 12: 32-37.

P. Keawram and W. Kanoksilapatham

Silpakorn U Science & Tech J Vol.7(2), 2013

Mexico. International Journal of Systematic and Evolutionary Microbiology 54: 169-174. Nam, G.-W., Lee, D.-W., Lee, H.-S., Lee, N.-J., Kim, B.-C., Choe, E.-A., Hwang, J.-K., Suhartono, M. T., and Pyun, Y.-R. (2002). Native-feather degradation by Fervidobacterium islandicum AW-1, a newly isolated keratinase-producing thermophilic anaerobe. Archives of Microbiology 178(6): 538-47. Nesbo, C. L., Bradnan, D. M., Adebusuyi, A., Dlutek, M., Petrus, A. K., Foght, J., Doolittle, W. F., and Noll, K. M. (2012) Mesotoga prima gen. nov., sp. nov., the first described mesophilic species of the Thermotogales. Extremophiles 16: 387-393. (Epub 2012 Mar 13) Patel, B. K. C., Morgan, H. W., and Daniel, R. M. (1985) Fervidobacterium nodosum gen. nov. and spec. nov., a new chemoorganotrophic, caldoactive, anaerobic bacterium. Archives of Microbiology 141: 63-69. Pasomsup, P., González, J. M., Portillo, M. C., Pongsapukdee, V., and Kanoksilapatham, W. (2011) Differentiation of a Hyperthermophilic Archaeon Pyrococcus sp. strain Pikanate 5017, by Arbitrarily Primed PCR. Silpakorn University Science and Technology Journal 5(1): 14-23. Ravot, G., Magot, M., Fardeau, M. L., Patel, B. K., Prensier, G., Egan, A., Garcia, J. L., and Ollivier, B. (1995) Thermotoga elfii sp. nov., a novel thermophilic bacterium from an African oil-producing well. International Journal of Systematic Bacteriology 45(2): 308-314. Stetter, K. O. (1996) Hyperthermophilic prokaryotes. FEMS Microbiology Reviews. 18: 149-158. Takahata, Y., Nishijima, M., Hoaki, T., and Maruyama, T. (2001) Thermotoga petrophila sp. nov. and Thermotoga naphthophila sp. nov., two hyperthermophilic bacteria from

Huber, R., Woese, C. R., Langworthy, T. A., Kristjansson, J. K., and Stetter, K. O. (1990) Fervidobacterium islandicum sp. nov., a new extremely thermophilic eubacterium belonging to the “Thermotogales”. Archives of Microbiology 154(2): 105-111. Jayasinghearachchi, H. S. and Lal, B. (2011) Oceanotoga teriensis gen. nov., sp. nov., a thermophilic bacterium isolated from offshore oil-producing wells. International Journal of Systematic and Evolutionary Microbiology 61: 554-560. Jeanthon, C., Reysenbach, A. L., L’Haridon, S., Gambacorta, A., Pace, N. R., Glénat, P., and Prieur, D. (1995) Thermotoga subterranea sp. nov., a new thermophilic bacterium isolated from a continental oil reservoir. Archives of Microbiology 164(2): 91-97. Kanoksilapatham, W., Pasomsup, P., Portillo, M. C. Keawram, P., and Gonzalez, J. M. (2012) Identification and Characterization of a Freshwater Pyrococcus sp. Strain PK 5017 and Identification of Pfu-Like IS Elements in Thermococcus sibiricus MM 739. International Journal of Biology 4(4): 11-22. L’Haridon, S., Miroshnichenko, M. L., Hippe, H., Fardeau, M. L., Bonch-Osmolovskaya, E. A., Stackebrandt, E. and Jeanthon, C. (2002) Petrotoga olearia sp. nov. and Petrotoga sibirica sp. nov., two thermophilic bacteria isolated from a continental petroleum reservoir in Western Siberia. International Journal of Systematic and Evolutionary Microbiology 52: 1715-1722. Miranda-Tello, E., Fardeau, M. L., Thomas, P., Ramirez, F., Casalot, L., Cayol, J. L., Garcia, J. L., and Ollivier, B. (2004) Petrotoga mexicana sp. nov., a novel thermophilic, anaerobic and xylanolytic bacterium isolated from an oil-producing well in the Gulf of

Silpakorn U Science & Tech J Vol.7(2), 2013

Diversity of Hyperthermophilic Bacteria

the Kubiki oil reservoir in Niigata, Japan. International Journal of Systematic and Evolutionary Microbiology 51(Pt 5): 19011909. Welsh, J. and McClelland, M. (1991) Genomic fingerprints produced by PCR with consensus tRNA gene primers. Nucleic Acids Research 19(4): 861-866. Wery, N., Lesongeur, F., Pignet, P., Derennes, V., Cambon-Bonavita, M. A., Godfroy, A., and Barbier, G. (2001) Marinitoga camini

gen. nov., sp. nov., a rod-shaped bacterium belonging to the order Thermotogales, isolated from a deep-sea hydrothermal vent. International Journal of Systematic and Evolutionary Microbiology 51: 495-504. Woese, C. R., Kandler, O., and Wheelis, M. L. (1990) Towards a natural system of organisms: proposal for the domains Archaea, Bacteria, and Eucarya. Proceedings of the National Academy of Sciences of the United States 87: 4576-4579.

Research Article A Study of Adsorption of an Organic Colouring Matter on Powdered Natural Plant Material Kiran V. Mehta Department of Chemistry, R. R. Mehta College of Science and C. L. Parikh College of Commerce, PALANPUR-385001, Dist.-Banaskantha, State-Gujarat (India) Corresponding author: E-mail: kiranvmehta@ymail.com Received April 22, 2013; Accepted July 1, 2013 Abstract

The present paper describes the study of the removal of an organic colouring matter (Astrazon Brilliant

Red 4G) (ABR) from the solution by the adsorbent prepared from wheat straw. To measure the efficiency of adsorption, different parameters like effect of pH, effect of initial dye concentration and effect of adsorbent amount with respect to time were studied. Spectrophotometric technique was used for the measurement of amount of dyes before and after adsorption process. The Langmuir adsorption isotherm model and Freundlich adsorption isotherm model were studied for this adsorption. The study reveals that wheat straw powder (WSP) can be used as an efficient adsorbent for the removal of ABR. Key Words: Adsorption; Environmental Pollution; Isotherms; Organic Colouring Matter; Removal. Introduction

Taneja (1994), Marmangne and Coste (1996),

Colour is an important feature of beauty.

Perkowski and Kos (2003), Rosario et al. (2002)

Colourants add colour to life and make life as

indicated that due to complex aromatic structures,

fine-looking as rainbow. Hence, the demand for

dyes are difficult to degrade and tend to persist in

colouring matter has experienced phenomenal

the environment and create severe water quality and

growth in the past some decades and the application

public health problems. This contaminated water

of these dyes has incessantly increased in many

affects the flora and fauna of the related region. As

industries. Industries like textile, rubber, drug,

a result, the environmental issues about the removal

paper, plastic, cosmetic, food and beverages use

of these pollutants are gaining much more attention

dye to colour their products. These coloured organic

in recent years.

substances are common water pollutants and they are

found in different quantities in industrial waste water.

eliminating dye from industrial effluents or coloured

This highly coloured water is discharged into nearby

water includes coagulation, chemical oxidation and

land, river or sea. Even at very low concentration,

biological treatment. However, these processes are

their presence in water is unquestionably visible

very pricey. The dyes from aqueous medium can be

and detrimental. Al-Ahmary (2013), Malik and

efficiently removed by adsorption process which

Silpakorn U Science & Tech J 7 (2) : 29-35, 2013

Conventional technologies employed for

ISSN 1905-9159

Silpakorn U Science & Tech J Vol.7(2), 2013

A Study of Adsorption of an Organic Colouring Matter

is a surface phenomenon. Dai (1998), Kannan and Sundram (2001) and Karaca et al. (2008) indicated in their work that a well known adsorbent, activated carbon can be prepared from carbonaceous material. As activated carbon is highly porous, it is a superb adsorbent but is very expensive. This has led to further research for the cheaper replacement of adsorbent material.

Extensive research has been directed to

the investigation of low cost materials as viable

Figure 1 Structure of colouring matter: Astrazon

substitutes for activated carbon. These materials

Brilliant Red 4G (ABR)

include palm fruit bunch (Nassar, 1997), cellulose based waste (Annadurai and Juang, 2002), compost

dye molecule is shown in Figure 1. Experimental

adsorbent (Lo Stuart and William, 2003), orange

solutions of desired concentrations were prepared

peel (Rajeshwari et al. (2003), beech saw dust

in distilled water.

(Bat Zias and Sindiras, 2004), peanut hull (Gong

Adsorbent

et al., 2005), rice husk (Ola et al., 2005), wheat

bran carbon (Ozer and Dursun, 2007), ginger waste

economy. Hence, agricultural wastes or agricultural

(Ahmad and Kumar, 2008), peach nut shells (Memon

by-products are easily available. India is a country

et al., 2009), fruit shell of Limonia Acidissima, etc.

producing and consuming a large amount of wheat.

(Torane et al., 2010).

Hence, dry wheat straws are easily available as an

To find efficient and environmentally friendly

agricultural by-product. Dry straws of wheat plants

adsorbents for the removal of various dyes from

used in the present study were collected from the

aqueous medium is a work of immense importance

local farms. They were washed four times with

to diminish the black dimension of industrialization.

distilled water to remove dust and water soluble

Like the other commercial colourants, use of

impurities, and dried till they became crisp. Then

colouring matter like Astrazon brilliant red 4G

it was sieved in the range of 45-50 mesh (US). The

(ABR) in textile industry creates pollution when

dried straws were further washed with distilled

untreated used coloured water is discharged. In the

water till the washings were free from turbidity.

present study, as a natural plant material, dry wheat

The powdered material obtained was dried at 105째C

straw powder (WSP) was used for the elimination

for 4-5 hours and placed in a dessicator. This WSP

of organic matter ABR from its aqueous solution.

was used as the adsorbent.

Agriculture is the backbone of the Indian

Adsorption Experiments

Materials and Methods

Adsorption of ABR on dry WSP was carried

Adsorbate

out. The effect of variations in pH values, effect

Astrazon brilliant red 4G (CAS No. 12217-

of initial concentration of adsorbate and effect of

48-0) was used as an adsorbate for the present work.

amount of adsorbent with respect to contact time

The dye has molecular formula C 23H 26N 3Cl

were studied. Simultaneously, all experiments were conducted with no adsorbent to ensure that

(molecular weight: 379.93 g/mol). lmax value of

adsorption was by wheat straw powder and not by

ABR in water is 514 nm. The structure of this

K. V. Mehta

Silpakorn U Science & Tech J Vol.7(2), 2013

the other factors. The process of adsorption was

temperature (33±1)°C for 120 min at various initial

studied by analyzing adsorption elimination of the

concentrations of ABR of 50, 100, 150, 200, 250

dye from the solution. The experimental mixture was

mg/L. The ABR removal was determined at 20, 40,

stirred on a rotary orbit shaker at 180 rpm. The study

60, 80, 100 and 120 min. The results of effect of

was performed according to the methods described

initial concentration of ABR on adsorption using

in the literature (Changwei et al., 2009; Dae-Hee,

WSP are shown in Table 1. It indicates that as initial

1999; Jusoh et al., 2004; Perineau et al., 1982;

concentration of ABR increases from 50 to 200

Sarkar and Bandyopadhyay, 2010). The parameters

mg/L, the % removal of ABR increases from 80.10

of Langmuir adsorption isotherm and Freundlich

to 85.01 respectively in the evaluation period of

Adsorption isotherm were also determined.

120 min.

The adsorption capacity, qe was calculated

The effect of amount of adsorbent on ABR

by:

removal was studied by varying the amount of

qe = V(Co – Ce) / W

adsorbent at 100, 300, 500, 700, 900 mg/100 ml

where qe is sorption capacity, V is the volume of the

respectively for 120 min. ABR concentration was

solution and W is the amount of the adsorbent, Co

kept 100 mg/l with the adsorption time of 120 min.

and Ce are initial and final adsorbate concentrations

The results of effect of amount of adsorbent on ABR

respectively.

removal are shown in Table 2. The results indicate

that the percentage of adsorption increases with

ABR removal percentage (%) = [(Co – Ce) /

the increase in amount of adsorbent. As adsorbent

Co] x 100

dose of WSP increases from 100 mg/100 ml to 900 Results and Discussion

mg/100 ml, significant increase in % ABR removal

was noticed i.e. from 62.11 to 89.14%.

For the determination of effect of pH, study

was carried out by taking 400 mg/100ml WSP at

(33±1) C for 150 min. The initial ABR concentration

Filipkowska et al., (2002) Kinnlburgh (1986)

was kept 100 ppm. The results of effect of pH are

and Longhinotti et al., (1998), it is clear that the

shown in Figure 2. The results indicate that at high

Langmuir and Freundlich adsorption isotherms

pH, removal of ABR was low, while at lower pH,

are most commonly employed models to depict

dye removal efficiency was found high.

the experimental data of adsorption isotherms. In

For determining effect of the initial dye

the present work, the adsorption data was analysed

concentration, the study was performed at fixed

with the help of Langmuir and Freundlich isotherm

adsorbent amount (400mg/100 ml) at room

models. The Langmuir isotherm is given by the

According to the work of Alau et al., (2010)

following equation: (Ce / qe) = (Ce / Q) + ( 1/Qb)

Here, Ce is equilibrium concentration of

adsorbate and qe is the amount of adsorbate adsorbed per unit mass of adsorbent. Q indicates Langmuir constant for the adsorption capacity while b is the Langmuir constant for energy of adsorption Figure 2 Effect of pH on removal efficiency

capacity. Ce /qe against Ce were plotted which gave 31

Silpakorn U Science & Tech J Vol.7(2), 2013

A Study of Adsorption of an Organic Colouring Matter

Table 1 Effect of initial concentration of ABR on adsorption using WSP Amount of adsorbent (WSP): 400 mg /100 ml, time: 120 min. Initial ABR concentration (mg/l)

50 100 150 200 250

18.01 23.32 25.65 24.05 22.88

% of ABR removal with time (min.) 40 60 80 100 39.95 40.79 46.26 44.85 44.10

62.85 55.30 61.10 57.80 60.20

77.25 75.10 75.50 74.95 76.14

80.14 82.70 82.95 83.30 81.10

120 80.10 82.65 83.24 85.01 82.11

Table 2 Effect of amount of adsorbent on ABR removal ABR concentration: 100 mg/L, time: 120 min. Adsorbent (WSP) dose (mg)

% of ABR removal with time(min.) 20

100

120

100

21.35

35.45

42.14

45.67

61.00

62.11

300

25.47

38.80

47.26

70.26

73.84

75.70

500

26.90

60.30

47.95

65.29

84.10

86.15

700

28.05

49.85

48.18

69.87

85.35

88.22

900

34.60

58.90

59.07

73.90

88.01

89.14

linear graph (Y = 0.0285 X + 0.520). The values of

that the isotherm is favourable for this kind of

Q and b were calculated on the basis of slope and

adsorption. The Langmuir adsorption is shown in

the intercept of the graph.

Figure 3. Values of related parameters of Langmuir

adsorption isotherm model for adsorption of ABR

Baseri, (2012) Hariharasuthan and

Nageswara, (2001) Mckay et al., (1984) and

on WSP are as below:

Tan et al., (2009) showed in their research work

that the essential characteristic of the Langmuir

coefficient(r) = 0.9996, RL = 0.1216.

Q = 35.0877, b = 0.0548, correlation

adsorption isotherm can be expressed in the terms of dimensionless equilibrium parameter RL. It is Ce/qe

given by the following equation: RL = 1/(1 + b Co)

Azraa (2012) showed that the value of RL

indicates the applicability of the chosen isotherm. The RL indicates the adsorption to be unfavourable (RL > 1), linear (RL = 1), favourable (0 < RL < 1) or

Figure 3 Langmuir isotherm for the adsorption

irreversible (RL = 0). The value of RL was obtained

0.1216 which is in the range of 0 to 1. It indicates 32

of ABR onto WSP

K. V. Mehta

Silpakorn U Science & Tech J Vol.7(2), 2013

significant increase in % ABR removal was noticed

Freundlich adsorption model can be expressed

i.e. from 62.11 to 89.14% respectively. So, it can be

by the following equation:

Log qe = (1/n) log Ce + log Kf

said that the percentage of adsorption increases with

Here, Kf and n are constants. The relation of

the increase in amount of the adsorbent. Thus, as a

log qe and log Ce was plotted results in a linear graph

natural plant material, dry WSP can be used as a

(Y = 0.5261 X + 0.545) indicating the adsorption

good, low cost adsorbent for the removal of the ABR

follows Freundlich adsorption isotherm. The value of

from aqueous solution. Langmuir and Freundlich

exponent ‘n’ signifies the feasibility of an adsorption

adsorption models can be applied for the adsorption

model while Kf denotes the ability of the adsorption.

isotherm of ABR on WSP.

The value of ‘n’ is found 1.9008 which is in the range of 1 to 10 indicating the adsorption model is

References

favourable. The Freundlich adsorption is shown in

Ahmad, R. and Kumar, R. (2008) Adsorption

Figure 4. Values of related parameters of Freundlich

Study of Patent Blue Vf Using Ginger

adsorption isotherm model for adsorption of ABR

Waste Material. Journal of the Iranian

on WSP are as below:

Chemical Society 1: 85-94. Al-Ahmary, K. M. (2013) Kinetics and

Kf = 3.508, correlation coefficient(r) =

Thermodyanamic Study of Malachite Green

0.9933, intercept = 0.5447.

Adsorption on Seeds of Dates. International Journal of Basic and Applied Sciences 2(1): Log qe

27-37. Alau K. K., Gimba C. E., and Kagbu J. A. (2010) Removal of Dyes from Aqueous Solution Using Neem (Azadirachta Indica) Husk as Activated Carbon, Archives of Applied

Log Ce

Science Research 2(5): 456-461.

Figure 4 Freundlich isotherm for the adsorption

Annadurai, G., Juang, R. S., and Lee, D. J. (2002)

of ABR onto WSP

Use of Cellulose Based Wastes for Adsorption of Dyes from Aqueous Solutions. Journal

Conclusion

of Hazardous Materials B-92: 263-274.

The amount of ABR adsorbed was found to

Azraa, A., Jain, K., Tong, K. S., Rozaini C., A., and

alter with pH and time. At lower pH, high removal

Tan, L. S. (2012) Equilibrium, Kinetic and

of the colourant from aqeous solution was observed.

Thermodynamic Studies on the Adsorption

The present study shows that of effect of initial

of Direct Dye onto a Novel Green Adsorbent

concentration of ABR on adsorption using WSP

Developed from Uncaria Gambir Extract.

indicates that as initial amount of ABR increased

Journal of Physical Science 23(1): 1-13.

from 50 to 200 mg/L, the % removal of ABR reached

Baseri, R. J., Palnisamy, P. N., and Sivakumar, P.

from 80.10 to 85.01 respectively in the evaluation

(2012) Comparative Studies of Direct Dye on

period of 2 hours. So, it is a good removal of the

Activated Carbon and Conducting Polymer

colourant from the solution. As adsorbent dose of

Composite. E-Journal of Chemistry 9(3):

WSP increased from 100 mg/0.1 L to 900 mg/0.1L,

1122-1134.

Silpakorn U Science & Tech J Vol.7(2), 2013

A Study of Adsorption of an Organic Colouring Matter

Bat Zias, F. A. and Sindiras, D. K. (2004) Dye

Kannan, N. and Sundram, M. M. (2001) Kinetic

Adsorption by Calcium Chloride Treated

and Mechanism of Removal of Methylene

Beech Sawdust in Batch and Fixed Bed

Blue by Adsorption on Various Carbons-A

Systems. Journal of Hazardous Materials

Comparative Study. Dyes and Pigments

114(1-3): 167-174.

1: 25-40.

Changwei, H., Jian, Long, L., Yin, Zhou, Mei, Li,

Karaca, S., Gurses, A., Aclkyldlz, M., and Ejder,

Feng, X., and Huiming, Li. (2009) Enhanced

M. (2008) Adsorption of Cationic Dye from

Removal of Methylene Blue from Aqueous

Aqueous Solutions by Activated Carbon.

Solution by Pummelo Peel Pretreated with

Microporous and Mesoporous Materials

Sodium Hydroxide. Journal of Health

115(3): 376-382.

Science 55(4): 619-624.

Kinnlburgh, D. G. (1986) General Purpose

Dae-Hee, Won-Seok, and Tai-II Y. (1999) Dyestuff

Adsorption Isotherms. Environmental

Wastewater Treatment Using Chemical

Science and Technology (J. Am. Chem. Soc.)

Oxidation, Physical Adsorption and Fixed

20(9): 895-896.

Bed Biofilm Process. Process Biochemistry

Lo Stuart, T., William, R. R., and Michael, A. C. (2003) Removal of Dissolved Textile Dyes

34: 429-439.

from Wastewater by a Compost Sorbent.

Dai, M. (1998) Mechanism of Adsorption for

Coloration Technology 119: 14-18.

Dyes on Activated Carbon. Journal of Colloid and Interface Science 198(1): 6-10.

Longhinotti, E., Pozza, F., Furlan, L., Sanchez,

Filipkowska, U., Klimiuk, E., Grabowski, S., and

M. D. N. D. M., Klug, M., Laranjeira, M.

Siedlecka, E. (2002) Adsorption of Reactive

C. M., and Favere, V. T. (1998) Adsorption

Dyes by Modified Chitin from Aqueous

of Anionic Dyes on the Biopolymer Chitin,

Solutions. Polish Journal of Environmental

Journal of the Brazilian Chemical Society

Studies 11(4): 315-323.

9(5): 435-440.

Gong, R., Ding, Y., Li, M., Yang C., Liu, H., and

Malik, A. and Taneja, U. (1994) Utilizing Flyash for

Sun, Y., (2005) Utilization of Powdered

Colour Removal of Dye Effluents. American

Peanut Hull as Biosorbent for Removal

Dyestuff Reporter October 20-26. Marmangne, O. and Coste, C. (1996) Colour

of Anionic Dyes from Aqueous Solution.

Removal from Textile Effluents, American

Dyes and Pigments 64: 187-192.

Dyestuff Reporter April 15-21.

Hariharasuthan, R. and Nageswara Rao, A. (2001) Comparative Studies of Sorel’s Cement on

Mckay, G., Blair, H. S., and Gardner, J. R. (1984)

Selected Dyes. Indian Journal of Science and

The Adsorption of Dyes onto Chitin in Fixed

Technology 4(4): 410 - 413.

Bed Columns and Batch Adsorbers. Journal

Jusoh, A., Tam, Y. K., Liew, A. G., Megat,

of Applied Polymer Science 29: 1499-1514.

Mohd., Noor, M. J., and Saed, K., (2004)

Memon, G. Z., Bhanger, M. I., and Akhtar, M. (2009)

Adsorption of Remazol Dye onto Granular

Peach-Nut Shells-An Effective and Low Cost

Activated Carbon in Fixed Bed : A Case

Adsorbent for the Removal of Endosulfan

Study of Red 3BS. International Journal of

from Aqueous Solutions. Pakistan Journal of

Engineering and Technology 1(1): 58-63.

Analytical and Environmental Chemistry 10(1-2): 14-18.

K. V. Mehta

Silpakorn U Science & Tech J Vol.7(2), 2013

Nassar, M. M. (1997) The Kinetics of Basic

Removal of Acid Violet 17(Acid Dye) from

Dye Removal Using Palm-Fruit Bunch.

Aqueous Solutions. Waste Management 21:

Adsorption Science and Technology 15: 609-

105-110. Rosario, L. C., Abel, G. E., and Marta, I. L. (2002)

617. Ola, A., Ahmed, E. N., Amany, E. S., and Azza,

Photodegradation of an Azo Dye of the

K. (2005) Use of Rice Husk for Adsorption

Textile Industry. Chemosphere 48: 393-399.

of Direct Dyes from Aqueous Solution: A

Sarkar, D. and Bandyopadhyay, A. (2010) Adsorptive

Case Study of Direct F. Scarlet. Egyptian

Mass Transport of Dye on Rice Husk Ash.

Journal of Aquatic Research 31(1): 1-11.

Journal of Water Resource and Protection 2: 424 - 431.

Ozer, A. and Dursun, G., (2007) Removal of Methylene Blue from Aqueous Solution by

Tan, I. A. W., Ahmad, A. L., and Hameed, B. H.

Dehydrated Wheat Bran Carbon. Journal

(2009) Adsorption Isotherms, Kinetics,

of Hazardous Materials 146(1-2): 262-269.

Thermodynamics and Adsorption Studies of

Perineau, F., Molinier, J., and Gaset, A. (1982)

2,4,6-Trichloro Phenol on Oil Palm Empty

Adsorption of Ionic Dyes onto Charred Plant

Fruit Bunch-Based Activated Carbon.

Material. Journal of Chemical Technolgy and

Journal of Hazardous Materials 164: 473-

Biotechnology 32: 749 -755.

482.

Perkowski, J. and Kos, L. (2003) Decolourization

Torane, R. C., Mundhe, K. S., Bhave, A. A., Kamble,

of Model Dye House Wastewater with

G. S., Kashalker, R. V., and Deshpande, N.

Advanced Oxidation Processes. Fibres and

R. (2010) Removal of Methylene Blue from

Textiles in Eastern Europe 11: 67-71.

Aqueous Solution Using Biosorbent. Der Pharma Chemica 2(3): 171-177.

Rajeshwari, S., Namasivayam, C., and Kadirvelu, K. (2003) Orange Peel as an Adsorbent in the

Research Article Test Case Based Selection for the Process of Software Maintenance Adtha Lawanna Department of Information Technology, Faculty of Science and Technology, Assumption University, Bangkok, Thailand Corresponding author. E-mail address: adtha@scitech.au.edu Received March 29, 2013; Accepted August 7, 2013

Abstract Software maintenance is the special process in the software-development life cycle. Particularly, the programmers have tried to reduce the size of testing and maintaining new software while fixing bugs is also realized. The large amounts of tests may cause time consuming, especially execution and operation. In response to this, many specialists propose the techniques for test case selection such as random selection and safe selection relying on the concept of regression testing. However, the ability of the new software is still required to be improved. Therefore, the test case based control-path is preferred to increase the performance of the program by creating and selecting the least test case as well as the faultless rate is preserved. Key Words: Software Development Life Cycle; Software Maintenance; Regression testing Introduction Due to the body knowledge of software engineering, this becomes an issue in developing programs. Up to now, most of the development teams are still creating new software for responding to the needs of users to support business objectives in their organizations (Carmel, 1995).Â In response to this, the software-development life cycle (SDLC) is powerful methodology that helps the programmers to produce the specific software, e.g., waterfall, iterative, prototyping, and spiral model (Larman and Basili, 2003). As we know, SDLC comprises the phase of user requirement, analysis, coding, testing, implementation, and maintenance (Cohen, 2010). User requirement is

Silpakorn U Science & Tech J 7 (2) : 36-45, 2013

the phase for gathering the needs and wants from users in details. After this is the step of analyzing the specific problem as well as preparing a good design, including data flow diagram, entity relationship, and database. Later, the programmers write the software, which is good or not depending on their skills and experiences. Next, is to monitor and correct the program, e.g., testing quality of software, entire system, and user satisfaction. Finally, in the process of SDLC is the software maintenance. The maintenance process is one of the most important phases in SDLC; particularly, it is designed to plan and control the new program in the entire system after adapting the existing software (Leau et al., 2012). The term maintenance

ISSN 1905-9159

A.Lawanna

Silpakorn U Science & Tech J Vol.7(2), 2013

includes fixing bugs, modifying, updating deleting, or adding some piece of the software. Modifying software may be required for adapting the system to the changes, e.g., technology, environment, or trend of customer life style. This paper considers those changes, including specification requirement, line of codes, and bugs. Sometime modification easily can be done, whereas there are few bugs occurred within the modified program. In general, bugs will be occurred, whenever the programmers do coding. On the other hand, modification may fail, in which the skills and experiences of the experts are involved e.g., the knowledge of programming languages, merging the difference codes, and the configuration (Chapin et al., 2001). As well as updating the software, the programmers may consider all functions and types of the program for selecting the important modules in order to improve its ability. The most difficulties of updating software are to change the previous programming language to the new and integrate the structure of the difference codes, including testing the entire system. The research area of software maintenance concerns the test suite selection, minimization, and prioritization. The techniques of test suite selection can be used to determine the numbers of the test cases from a test suite (Harrold et al., 1993 ; Harrold, 1999). Particularly, any test suite contains a set of test case, which is created relying on the specific factors e.g., requirements, codes, and bugs. More specifically, another purpose of software maintenance is to preserve the faultless within the changed program (Niessink, 2000). Those factors affect the entire software system in terms of executing and running the modified software. The execution time is a major problem when all test cases are audited as well as the running time (Musa, 1993). As we know, the higher numbers

of the test cases often show the better abilities of the new software, whereas the faults are small. However, the maximum numbers of the test cases increase the execution time. Therefore, many researches propose the techniques that select the minimum numbers of test cases as well as fixing bugs within the new software. This paper presents the technique that can solve the remained problems by selecting the lower numbers of test cases while the faultless rate is preserved than the traditional technique. The proposed model concerns the subject programs, specifically used in the area of selecting test cases and decreasing the faults of the software. In addition, this paper shows some of the traditional techniques, which are used to compare their abilities. Materials and Methods Data set Preparing the experiment is one of the most important methods. Accordingly, the data set is required. In Table 1, the seven subject programs are required whereas the program name, numbers of function (F), lines of code (C), faulty versions (V) and the test suites (T) are available (Rothermel and Harrld, 1998). To manage a test suite and automate test execution, a test database management system is created, and playback tools are captured (Rothermel, 1996). Those subject programs are written by the developers of the Siemens suite of programs with manually fixing bugs or faults. The artifacts of all seven programs have consequently, been revised and extended by other agents. These programs are preferred because of the development of the related artifacts as well as the historical significance. Numerous high-quality experimental software engineering researchers have used the Siemens suite (Ostrand, 1998).

Silpakorn U Science & Tech J Vol.7(2), 2013

Test Case Based Selection for the Process of Software Maintenance

Table 1 The Subject programs Program Name

Numbers of function(F)

Line of codes(C)

Faulty Versions (V)

Test Suite (T)

print-tokens

402

4,130

print-tokens2

483

4,115

replace

516

5,542

schedule

299

2,650

schedule2

297

2,710

tcas

148

1,608

totinfo

346

1,052

Regression Testing (RT) Regression testing is the method of testing changes within software or programs to ensure that the existing system still work with the new changes (Agrawal et al., 1993). Regression testing is a basic part of the SDLC, especially in the software maintenance. For the large companies, RT is done by software testing specialists or programmers. The typical steps of RT are described as follows: (1) Select the test cases from a test suite. A test suite is the set of the test cases, which can be constructed automatically by the test case generator. T = {t1 , t 2 , t 3 ,..., t n } (1)

program. If the selected test cases by (2) cannot cover all the specifics requirements, then the new test cases should be chosen for fixing this problem.

(3)

Therefore, the total selected test cases equal ∗

∗∗

t +t These test cases form what becomes the test bucket. Before releasing a new version of a software product, the old test cases are also run against the modified version in order to make sure that all the exist capabilities still run. The reason that they might not work is because modifying or adding new code to a program can easily produce bugs into code that may not have intended to be made. Test department coders do program test scenarios and exercises that will test new modules of code after they have been written. Researchers have tried to perform regression testing more efficient and more effective by preparing regression test selection (RTS) techniques, but many problem remain, such as: RTS techniques may save time and money, however they sometimes may select most or all of the original test cases (Leung and White, 1991). Therefore, specialists using RTS techniques can find themselves worse off for having done so

Where: T is a test suite and t is the test case. (2) Test the program (P) with the selected test cases. In general, a test suite contains huge amounts of the test cases. Therefore, the developers select some of the test cases for the process of software maintenance, e.g., bugs and run time.

t ∗∗ = {t1 , t 2 , t 3 ,..., t v }

t ∗ = {t1 , t 2 , t 3 ,..., t m } (2)

Where: t ∗ is a set of selected test cases regarding the specifications requirements from users and developers. (3) If necessary, create new test cases for the

A.Lawanna

Silpakorn U Science & Tech J Vol.7(2), 2013

the least of bugs are performed. This technique can provide the better results in lower cost and timeless in the process of execution and testing (Wong et al., 1997). The steps of SS are explained as follows: (1) Due to using the RTS, a test suit is given. (2) Create a control flow graph (CFG) for the program. In a program, control flows are created from variables and procedures, e.g. such as if- and while-constructs in the programming language. It is a representation of its possible control flows through the whole program. Particularly, all nodes correspond to statements and decisions, including edges are used to represent the flow of control in a code. Statement coverage is constructed in a test

(Ball, 1998). Testing time is often limited, e.g., must be finished overnight. RTS techniques do not focus such problems and, hence, can select more test cases than can be work. RTS techniques can maximize the average regression testing ability rather than optimize aggregate ability over many parts of testing software. Random Selection (RS) Random Selection is the technique that is created after the Testing All-Selection (AS) is applied. The major benefit of AS is the minimum numbers of faultless rate. However, it may cause a big problem of time consuming. Therefore, many development turns to the RS because it is simplest and to avoid the high cost, including timeless (Grave et al., 2001). The steps of RS are explained as follows: (1) Due to using the RTS, a test suite is given. (2) Randomly select the test cases from a test suite. This step can be done by a spreadsheet Microsoft ExcelTM. For example, if there is a test suite (T), the numbers of the selected test cases can be computed by “=T(RAND())”. Note: RS technique gives the least execution time for module testing, but it may not guarantee the ability of the program in terms of producing the new bugs whereas the entire software is not tested. More details can be found in the article of Grave and team, 2001. Safe Selection (SS) Safe Selection is the technique an efficient regression selection (Hutchins et al., 1994; Rothermel and Harrld, 1996; Rothermel and Harrold, 1997; Rothermel and Harrold, 1998). It is one of the regression test selections implemented as a tool called DejaVu. Specifically, this technique provides smaller numbers of test cases compared with AS, and RS. Another reason of using the SS is

suite that can execute every statement at least once of a whole program. (3) Test execution profiles and choose all test cases in a test suite that, when executed through the program. (4) Exercise the program at least on statement that is deleted from the program, or that, when executed on the modified version. (5) Exercise the modified program at least on statement that becomes a new or modified in the latest version. The statement that does not exist in the program cannot be executed. Therefore, the selected test cases can be provided by exercising the program or the modified version, which is created to be safe. The conceptual overview of the proposed model namely Lawanna Selection (LS) The activities in the process of software maintenance and Lawanna Selection are shown in Figure 1. The details of the whole steps are described as follows; (1) After the process of softwaredevelopment life cycle is reached, the software will be released to the users.

Silpakorn U Science & Tech J Vol.7(2), 2013

Test Case Based Selection for the Process of Software Maintenance

(2) When the time goes by, users may need the modified software. Therefore, the development team needs the specification requirements from the users in details. (3) After this, the programmers will modify the previous program regarding (2). (4) Testers test the modified program, e.g., checking inputs, functions, and outputs of the code, including fixing bugs. (5) When faults or bugs are produced, it is necessary to redo (4) again. (6) Check the side effect of a whole program, e.g., the relationship of variables, functions, and the expected results such as the ability of running the program, execution time, and user acceptance. Particularly, this activity refers using the regression testing. (7) A test suite is given by (6), there are the large amounts of the test cases are generated. This causes the complexity of testing the software, the consequence of corresponding failures, difficulty of solving errors, and debugging the test cases. In this research, the LS is proposed to solve the problem about the increasing size of a test suite by selecting few test cases, which can preserve the competency of the whole program. However, one problem can be occurred in LS, which is producing the irrelevant selected test cases. These test cases cannot be run properly or there are bugs found. Therefore, to fix this problem is necessary to regenerate the test cases again. Besides this, LS technique needs to use the outcome of the regression testing by realizing the major variables, which are the value of F and C from the general subject program. Come to this point, the modified program will be released to the users, whereas the whole processes are done.

Deploy software Get requirements Modify code Unsolved problem Errors

Test the modified program Create regression test Irrelevant test cases Apply Lawanna Selection Release new software version

Figure 1 The process of software maintenance and Lawanna Selection Lawanna Selection (LS) concerns the relationship of the numbers of function (F), line of codes (C), and the faulty versions (V). Eq. 4 to Eq. 6 are shown as follows;

F = {1,2,3,...,n} (4)

C = {1,2,3,...,m} (5)

V = {1,2,3,...,r} (6)

In particular, not only the value of F, C, and V are required by developers, but it includes the user requirements and test case generator. This is because they affect the size and quality of a test suite and the competency of a whole program.

A.Lawanna

Silpakorn U Science & Tech J Vol.7(2), 2013

A set of the selected test cases is generated and located in a test suite. One of the main objectives is to get the small test cases by avoiding the problem of run time. Next, LS can prepare the higher reduction rate, including providing the least faults or bugs in the modified program. The conceptaul model of Lawanna Selection (LS) The most important issue in the process of software maintenance is to preserve faultless rate of the minimal selected test cases in order to avoid the bugs that can be occurred. In response to this, the model of LS is proposed to build the concept of selecting the relevant test cases in any test suite of the program through the process of maintaining software.

The subject program

techniques involve creating a test suite. The reason is the complexity of a program, which combines all conditions of writing source code, e.g., user requirements, numbers of function, inputs, and expected outputs. Therefore, many researches import a given test suite which is automatically created by a specific test case generator. In the LS, a test suite also can be generated by the software named Reactis Tester. The test cases can be generated by importing inputs and clearly steps of testing in order to provide the appropriate output. A relevant test case will give “pass” not “fail” at the expected output. The most complicate part of LS is to create the test case path or control-path. Accordingly, the control-path shows the steps constructed in each test case. The first assumption by applying LS is that the generated test cases have their own steps of dealing with the different inputs for checking the outputs (pass or fail). The second assumption is the outputs of all generated test cases are “pass”. The reason is that 100% coverage is required. If “fail” is found, then that test case will be rejected. Besides this, one of the most important steps of the LS is to select the appropriate test cases. Of course, each test case may take the same or different steps for testing the specific inputs. For LS, it needs the shortest steps to be the representative. Surely, the expectation of proposing LS is to select small amounts of the test cases with 100 % coverage to avoid technical errors and keep the specification requirements. The experimental steps of LS (1) Define a given test suite (TS) This step is created in order to define a test suite, which can be generated by Reactis Tester. A test suite will be constructed by executing the subject program, which the input and output values are recorded at each step. (2) Generate the test cases. The test case template is created as shown in

Define a test suite

Generate the test case

Create the test case path The set of selected test cases

Select the test case

Figure 2 The Lawanna Selection (LS) Figure 2 shows the conceptual model of the Lawanna Selection (LS). The process of LS starts when the users require the updated program for their purposes. Accordingly, the requirements relate directly to the numbers of function (F), line of codes (C), and numbers of bug (B). The general problem of LS and many traditional selection

Silpakorn U Science & Tech J Vol.7(2), 2013

Test Case Based Selection for the Process of Software Maintenance

in order to avoid producing the irrelevant test cases (the coverage value is less than 100%). For example, if there are x steps in test case (1), which is constructed to handle y inputs, and all y outputs can be produced and reported. This means the test case (1) has 100% coverage. Therefore, test case (1) is usable, if it is not 100%, it will be rejected from a test suite.

Table 2. It comprises the test case number, Input, output, and step of testing. The test cases can be generated regarding the percentage of covering the specifics requirements. Due to defining a test suite, it gives a test suite that contains the test cases with different % coverage. In the experiment, only the test cases with 100% coverage will be generated Table 2 The test case template Test Case (No.)

Input (1)

Input (2)

…

Input (y)

Output (1)

Output (2)

…

Output (y)

Step (1) Step (2) … Step (x)

(3) Create the control-path. Assume that all inputs and output of a test case are active. The numbers of step are realized for creating the control-path. The good control-path will show the least steps of testing. On the other side, if the control-path (H) takes many steps, the control-path may not be appropriate. The controlpath can be written as Eq. 7.

end if Therefore, a set of the control-path, H = {H1, H2, H3,...,Hx} (4) Select the appropriate test cases. According to the algorithm of creating

H x = Step ( x ) (7)

if H min = {t (1), t (10), t ( 45), t (100), t (124)} , then

control-path, a set of H min is constructed. In fact, the result of H min can show the test cases number that has the minimum steps. For example, t * = {t (1), t (10), t ( 45), t (100), t (124)} . Therefore, the

Where; x is the number of step for testing a test case. The shortest control-path has minimum steps of testing the test case. Accordingly, a test case with the minimum step is required. Algorithm of creating the control-path

numbers of the selected test cases equal to 5 or t* = 5. (5) Find the reduction rate (RR) can be written as Eq. 8. ∗ RR = T − t (8) T

if H = Step (1) then

The reduction rate is the ratio of the remained test cases and a test suite. (6) Determine the faultless rate (FR) can be written as Eq. 9.

create H 1 else if H = Step ( 2) then create H 2

else if H = Step (x ) then

create H x

∗ FR = 1 − t −∗ B (9) t

A.Lawanna

Silpakorn U Science & Tech J Vol.7(2), 2013

The faultless rate refers to the possibility of finding bugs in a set of the selected test cases. Eq. 9 is required for evaluating the ability of the comparative studies, e.g., RS, SS, and LS. However, the value of B is assumed to be 1 for the computation. This means every selected test case should avoid the numbers of bug. In worst situation, the only bug is allowed in a set of selected test case. However, the bug needs to be fixed by the programmer before the deployment.

example of computing the reduction rate of LS on the program named print-token is shown as; . RR = 4130 â&#x2C6;&#x2019; 67 = 0.9838 4130

The results of finding the reduction rate due to the same computation of RS and SS are 0.9075 and 0,9230 respectively. According to this result, the reduction rate of RS is lower than SS and LS. This is because the numbers of the selected test cases are higher than others. Therefore, the second contribution of LS is to provide the higher reduction rate compared with the traditional techniques.

Results and Discussions In facts, there are many selection techniques are developed for improving the performance of eliminating the size of a test suite. In this paper, there are three comparative studies, which are RS, SS, and LS. This is because RS is the well-known and simplest technique that is used in the part of evaluating the ability of the comparative studies. Another is SS, which is the powerful technique in the field software maintenance. As we can see, Table 3 shows the numbers of the selected test cases by RS, SS, and LS. The numbers of the selected test cases by LS are lower than others. Therefore, one of the benefits of using LS is to provide the smallest size of a test suite.

Table 4 Reduction Rate of the comparative studies Program Name

print-tokens

382

318

print-okens2

299

389

replace

426

398

schedule

483

225

schedule2

234

tcas

203

totinfo

214

199

148

print-tokens

0.9075

0.9230

0.9838

print-okens2

0.9273

0.9055

0.9815

replace

0.9231

0.9282

0.9868

schedule

0.8177

0.9151

0.9811

schedule2

0.9790

0.9137

0.9793

tcas

0.8738

0.9484

0.9689

totinfo

0.7966

0.8108

0.8593

Table 5 shows the value of faultless rate of the comparative studies, which can be calculated by using Eq. 9. The example of computing the faultless rate at least one bug found of LS on the program named print-token is shown as;

Table 3 The numbers selected test cases of the comparative studies Program Name

FR = 1 â&#x2C6;&#x2019; 67 â&#x2C6;&#x2019; 1 = 0.0149 . 67

The results of finding the faultless rate due to the same computation of RS and SS are 0.0026 and 0,0031 respectively. Regarding to Table 5, we can summarize that the results of finding the faultless rate of LS are higher than others. This means that the probability of producing bugs in a whole set of the selected test cases by using LS is less than RS and SS.

Table 4 presents the reduction rate of all programs that can be computed by using Eq. 8. The

Silpakorn U Science & Tech J Vol.7(2), 2013

Test Case Based Selection for the Process of Software Maintenance

Table 5 Faultless Rate of the comparative studies Program Name

print-tokens

0.0026

0.0031

0.0149

print-okens2

0.0033

0.0026

0.0132

replace

0.0023

0.0025

0.0137

schedule

0.0021

0.0044

0.0200

schedule2

0.0175

0.0043

0.0179

tcas

0.0049

0.0120

0.0200

totinfo

0.0047

0.0050

0.0068

Chapin, N., Hale, J. E, Ramil, J. F., and Tan, W. (2001) Types of software evolution and software maintenance. Journal of Software Maintenance and Evolution 13(1): 3-30. Cohen, S. (2010) A Software System Development Life Cycle Model for Improved Stakeholdersâ&#x20AC;&#x2122; Communication and Collaboration. International. Journal of Computers, Communications and Control 5(1): 20-24. Harrold, M. J., Gupta, R., and Soffa, M. L. (1993) A methodology for controlling the size of a test suite. ACM Transactions on Software Engineering and Methodology 2(3): 270285. Harrold, M. J. (1999) Testing Evolving Software.

Conclusion The Lawanna Selection model is the alternative technique for the process of software maintenance by using the concept of regression test selection. It provides the process of selecting the minimum numbers of the test cases while the performance of the program is preserved. Particularly, when compare LS with the traditional techniques such as RS and SS. There are three benefits of using LS. First, the size of the selected test cases by using the LS is smaller than applying the RS and SS. Second, it gives the higher reduction rate than the traditional techniques. Third, LS gives lower numbers of producing the new bugs than RS and SS technique.

Journal of Systems and Software 47(2): 173181. Hutchins, M., Foster, H., Goradia, T., and Ostrand, T. (1994) Experiments on the effectiveness of dataflow- and control flow-based test adequacy criteria. International of Software Engineering 16(1): 191-200. Larman, C. and Basili, V. R. (2003) Iterative and Incremental Development: A Brief History. Journal of Computer 36(6): 47-56. Leau, Y. B., Loo W. K., Tham, W. Y., and Tan S. F. (2012) Software Development Life Cycle AGILE vs Traditional Approaches. International Conference on Information and Network Technology 37(1): 162-167. Leung, H. K. N. and White, L. J. (1991) A cost model to compare regression test strategies. In Proceedings of the Conference on Software Maintenance, 201-208. Niessink, F. and Van, V. H. (2000) Software maintenance from a service perspective. Journal of Software Maintenance and Evolution : Research and Practice 12(1): 103-120.

Reference Agrawal, H., Horgan, J., Krauser, E., and Londo, S. (1993) Incremental regression testing. In Proceedings of the Conference on Software Maintenance, 348-357. Ball, T. (1998) The limit of control-flow analysis for regression testing. In International Symposium on Software Testing and Analysis, 143-242. Carmel, E. (1995) Cycle Time in Packaged Software Firms. Journal of Product Innovation 12:110-123.

A.Lawanna

Silpakorn U Science & Tech J Vol.7(2), 2013

Musa, J. (1993) Operational proďŹ les in software reliability engineering. IEEE Software 10(2): 14-32. Ostrand, T. and Balcer, M. (1988) The categorypartition method for specifying and generating functional tests. ACM Transactions on Software Engineering and Methodology 31(6): 676 - 686.Â Grave, T. D., Harrold M. J., Kim, J. M., Porter, A., and Rothermel, G. (2001) An empirical comparison of regression test selection techniques. ACM Transactions on Software Engineering and Methodology 10(2): 184-208. Rothermel, G. (1996) Efficient, effective regression testing using safe test selection techniques. Technical Report Clemson University 96-101. Rothermel, G. and Harrld, M. J. (1996) Analyzing regression test selection techniques. IEEE Transaction of Software Engineering. 22(8): 529-551.

Rothermel, G. and Harrld, M. J. (1997) A safe, efficient regression test selection technique. ACM Transactin of Software Engineering 6(2): 173-210. Rothermel, G. and Harrld, M. J. (1998) Empirical studies of a safe regression test selection technique. IEEE Transaction of Software Engineering 24(6): 401-419. Wong, W., Horgan, R., London, S., and Mathur A. (1997) A study of effective regression testing in practice. In 8th International Symposium on Software Reliability Engineering, 264275. Wong, W. E., Horgan, J. R., Mathur, A. P., and Pasquini, A. (1997) Test set size minimization and fault detection effectiveness: A case study in a space application. In Proceedings of the 21st Annual International Conference on Computer Software and Applications. 522-528.

ISSN 1905-9159

Silpakorn University

Science and Technology Journal Volume 7 Number 2 (July-December) 2013

Double Bootstrap-t One-sided ConďŹ dence Interval for Population Variance of Skewed Distributions Wararit Panichkitkosolkul

A Study of Adsorption of an Organic Colouring Matter on Powdered Natural Plant Material Kiran V. Mehta

Test Case Based Selection for the Process of Software Maintenance The subject program

Define a test suite

Generate the test case

Create the test case path

http://www.surdi.su.ac.th http://www.journal.su.ac.th http://www.tci-thaijo.org/index.php/sustj

The set of selected test cases

Select the test case

Adtha Lawanna