Shyamal D. Peddada, PhD

Professor and Chair, Biostatistics


7126 Public Health, 130 DeSoto Street, Pittsburgh, PA 15261
R-znvy: fqc92@cvgg.rqh
Primary Phone: 967-838-9357
Web site:

Personal Statement

In this research program, we develop broadly applicable statistical methods that are motivated by applications in biomedical sciences, such as microbiome, genomics, high throughput screening assays, toxicology, oscillatory systems such as circadian clock and cell-cycle, etc. Two important features of the methods developed in our research program are (a) they take into account the underlying structure or constraints in the scientific problem, and (b) they are applicable to high dimensional data. We also collaborate extensively with researchers on a wide range of projects where we often use the methods developed in this research program.  We are currently collaborating on a variety of projects relating to human microbiome, toxicology, and genomics. 


1983 | University of Pittsburgh, Pittsburgh, PA | PhD, Statistics
1981 | University of Pittsburgh, Pittsburgh, PA | MA, Statistics
1980 | Indian Agricultural Research Institute, New Delhi, India | MSc, Agricultural Statistics
1977 | University of Delhi, Delhi, India | BSc (HONS), Mathematics

Selected Publications


A. Theory, Methods and Applications


Peddada SD. Seasonal change in the gut: The gut mictobiome of Hadza hunter-gatherers changes with the season. Science 357 (6353), 754-755 (2017). [Invited Perspective]
Weiss S, Xu ZZ, Peddada SD, Amir A, Bittinger K, Gonzalez A, Lozupone C, Zaneveld JR, Vázquez-Baeza Y, Birmingham A, Knight R. Normalization and microbial differential abundance strategies depend upon data characteristics. Microbiome. DOI: 10.1186/s40168-017-0237-y. (2017).
Davidov O, Jelsema C, Peddada SD. Testing for inequality constraints in singular models by trimming or winsorizing the variance matrix. J. Amer. Statist. Assocn. (2017), in press. 
Kaul A, Davidov O, Peddada SD. Structural zeros in high-dimensional data with applications to microbiome studies. Biostatistics 18, 422-433 (2017).
Mandal M , Godfrey K , McDonald D , van Treuren W , Bjørnholt JV , Midvedt T, Moen B , Rudi K , Knight R , Peddada SD, Eggesbø M. Fat and vitamin intakes during pregnancy have stronger relations with a pro-inflammatory maternal microbiota than does carbohydrate intake. Microbiome 4(1):55 DOI: 10.1186/s40168-016-0200-3 (2016).
Larriba Y, Rueda C, Fernandez MA, Peddada SD. Order restricted inference for oscillatory systems for detecting rhythmic signals. Nucleic Acids Research doi:  10.1093/nar/gkw771 (2016).
Jelsema C, Peddada SD. An R Package for Linear Mixed Effects Models under Inequality Constraints. Journal of Statistical Software doi: 10.18637/jss.v075.i01 (2016). 
Mandal S, Van Treuren W, White RA, Eggesbø M, Knight R, Peddada SD. Analysis of composition of microbiomes: a novel method for studying microbial composition. Microbial Ecology in Health and Disease 26, 1 – 7 (2015).
Davidov O, Peddada SD. Testing for the Multivariate Stochastic Order among Ordered Experimental Groups with Application to Dose-Response Studies. Biometrics 69, 982-990 (2013).
White RA, Bjørnholt J, Baird DD, Midtvedt T, Harris JR, Pagano M, Hide W, Rudi K, Moen B, Iszatt N, Peddada SD, Eggesbø M. Novel Developmental Analyses Identify Longitudinal Patterns of Early Gut Microbiota that Affect Infant Growth. PLOS Computational Biology 9(5):e1003042 (2013)
Davidov O, Peddada SD. The linear stochastic order and directed inference for multivariate ordered distributions. Annals of Statistics 41(1):1-40 (2013).
Fernandez M, Rueda C, Peddada SD. Identification of a core set of signature cell-cycle genes whose relative order of time to peak expression is conserved across species. Nucleic Acids Research, 40,2823-32, doi: 10.1093/nar/gkr1077 (2012).
Davidov O, Peddada SD. Order restricted inference for multivariate binary data with application to toxicology. J. Amer. Statist. Assoc., 106, 1394-1404 (2011).
Guo W, Sarkar SK, Peddada, SD. Controlling False Discoveries in Multidimensional Directional Decisions, with Applications to Gene Expression Data on Ordered Categories. Biometrics, 66, 485 – 492 (2010).
Peddada SD, Laughlin S, Miner K, Guyon J-P, Haneke K, Vahdat HL, Semelka RC, Kowalik A, Armao D, Davis D, Baird DD. Growth of uterine leiomyomata among premenopausal black and white women.  Proc. National Acad. Sci., 105, 19887-19892.
Peddada, SD, Dinse, G and Kissling, G. Incorporating Historical Control Data When Comparing Tumor Incidence Rates. J. Amer. Statist. Assoc., 102, 1212-1220 (2007).
Peddada SD, Lobenhofer L, Li L, Afshari C, Weinberg C and Umbach D. Gene Selection and Clustering for Time-course and Dose-response Microarray Experiments using Order-restricted Inference. Bioinformatics, 19, 834-841 (2003).
Hwang JTG and Peddada SD. Confidence Interval Estimation Subject to Order Restrictions. Annals of Statistics, 22, 67-93 (1994).


B. Collaborative Research


Hill-Burns EM , Debelius JW, Morton JT, Wissemann WT, Lewis MR , Wallen ZD, Peddada SD, Factor SA, Molho E, Zabetian CP, Knight R, Payami H. Parkinson's disease and Parkinson's disease medications have distinct signatures of the gut microbiome. Movement disorders: official journal of the Movement Disorder Society 32(5), 739-749 (2017).
Reese SE, Zhao S, Wu M, Joubert BR, Parr CH, Haberg S, Ueland PM, Nilsen RM, Midttun O, Vollset SE, Peddada SD, Nystad W, London SJ. DNA Methylation Score as a Biomarker in Newborns for Sustained Maternal Smoking during Pregnancy. Environmental health perspectives doi:10.1289/EHP333 (2017).
Rebera SO, Siebler PH, Donner NC, Morton JT, Smith DG, Kopelman JM, Lowe KR, Campbell K, Fox JH, Hassell JE, Greenwood BN, Janscha C, Lechner A, Uschold-Schmidt N, Füchsl AM, Langgartner D, Walker FR, Hale MW, Perez GL, Treuren WV, González A, Halweg-Edwards AL, Fleshner M, Raison CL, Rook GA, Peddada SD, Knight R, Lowry CA. Immunization with a heat-killed preparation of the environmental bacterium Mycobacterium vaccae promotes stress resilience in mice. Proc. National Acad. Sci. doi: 10.1073/pnas.1600324113 (2016).
Joubert B, den Dekker H, Felix J, Bohlin J, Ligthart S, Beckett E, Tiemeier H, van Meurs J, Uitterlinden A, Hofman A, Haberg S, Reese S, Peters M, Andreassen B, Steegers E, Nilsen R, Vollset S, Midttun O, Ueland P, Franco O, Dehghan A, de Jongste J, Wu M, Wang T, Peddada SD, Jaddoe V, Nystad W, Duijts L, London S. Maternal plasma folate impacts differential DNA methylation in an epigenome-wide meta-analysis of newborns. Nature Commun. doi: 10.1038/ncomms10577 (2016).
Yamashita, H, Hoenerhoff, MJ, Peddada, SD, Sills, RC, Pandiri, AR. Chemical Exacerbation of Light-induced Retinal Degeneration in F344/N Rats in National Toxicology Program Rodent Bioassays. Toxicologic pathology 44(6), 892-903 (2016).


Selected Software


The following software was developed in this research program and is freely available to download from


Order Restricted Inference for Oscillatory Systems (ORIOS) for Detecting Rhythmic Signals
(R code developed by Yolanda Larriba, University of Valladolid, Valladolid, Spain).
ORIOS is a model free order restricted inference based algorithm that detects rhythmic components (e.g. transcripts or genes) s participating in oscillatory systems such as the circadian clock. Although this software can be used for any oscillatory data, for simplicity of description, throughout this file we shall use the term “circadian clock data” rather than “oscillatory data” and “genes” in place of “components” of an oscillatory system.  The strength of model free methodology such as the order restricted inference is that, instead of using a mathematical model to describe the shape or pattern of expression, it uses mathematical inequalities to describe patterns.  Thus it is robust and is not limited by any mathematical model that may be rigid and restricted. ORIOS not only identifies rhythmic genes, it also classifies them into four typical classes of genes, called cyclical, quasi cyclical, non-flat and non-periodic, and flat, according to its signal shape.  Cyclical and quasi cyclical genes are declared as rhythmic, while non-flat and non-periodic, and flat are declared as non-rhythmic genes. Compared to some commonly used rhythmicity detection algorithms in, ORIOS has substantially higher power to detect true rhythmic genes, while also declaring substantially fewer non-rhythmic genes as rhythmic.


• Larriba Y, Rueda C, Fernandez MA, Peddada SD. Order restricted inference for oscillatorysystems for detecting rhythmic signals. Nucleic Acids Research doi: 10.1093/nar/gkw771 (2016).


Analysis of Compositional Microbiomes (ANCOM) data

(R-code developed by Dr. Siddhartha Mandal, Norwegian Institute of Public Health and the Shiny app developed by Dr. Casey Jelsema, Biostatistics and Computational Biology Branch, NIEHS)
This R package is designed for comparing the abundance of individual taxa in two populations using log-ratios of abundance. This software is based on the ANCOM methodology developed in Mandal et al. (2015).




• Mandal S, Van Treuren W, White RA, Eggesbø M, Knight R, and Peddada SD. Analysis of composition of microbiomes: a novel method for studying microbial composition. Microbial Ecology in Health and Disease, 26, 1 – 7 (2015).


R Code for Estimating of Global Relative Order of Peak Expression Satisfied by a Set of Oscillatory Genes(28KB)

(Programmed by Ms. Sandra Barragán, University of Valladolid, Spain):
For a given collection of oscillatory genes (e.g. cell-cycle genes or circadian clock genes) with phase angles estimated from multiple experiments, in this software we estimate the relative order of peak expression among the genes. It contains 2 functions written in R, called Aggregation of Circular Orders (ACO), which is based on a solution to the traveling salesman problem, and Circular Local Minimization (CLM) algorithm which is used to smooth the solution obtained from ACO. To run these programs the user should first download the companion R package called \emph{isocir} from CRAN .




• Barragán S, Rueda C, Fernández MA, Peddada SD (2015) Determination of Temporal Order among the Components of an Oscillatory System. PLoS ONE 10(7): e0124842.


ORIOGEN 4.01 - Order Restricted Inference for Ordered Gene Expression and Multiple Pairwise Comparisons:

This software is designed for comparing two or more experimental groups. There are two options available within software, with one used for analyzing ordered experimental conditions (e.g. time, dose, tumor stages, etc.). Under this option, the software can handle an independent sample case, as well as a dependent sample case (e.g. repeated measurements). The residual bootstrap methodology used in this software is robust to any underlying dependence structure. The method controls the FDR at the desired level. The second option is suitable for pairwise comparisons and is not limited to ordered experimental conditions. Thus, for any given design, the second option allows one to make all desired pairwise comparisons among the experimental groups. In addition it allows one to make directional inferences (such as up or down regulated genes etc.). The method controls for the overall mixed directional false discovery rates (mdFDR).


• Guo W, Peddada SD. Adaptive Choice of the Number of Bootstrap Samples in Large Scale Multiple Testing. Statistical Applications in Genetics and Molecular Biology, 7 (1), Art. 13 (2008).
• Peddada SD, Harris S, Zajd J, Harvey E. ORIOGEN: Order Restricted Inference for Ordered Gene Expression data. Bioinformatics, 21, 3933-3934 (2005).
• Peddada SD, Lobenhofer L, Li L, Afshari C, Weinberg C, Umbach D. Gene selection and clustering for time-course and dose-response microarray experiments using order-restricted inference. Bioinformatics, 19, 834-841 (2003).
Constrained Linear Mixed Effects (CLME) for analyzing mixed and fixed models under inequality constraints.
(Programmed by Dr. Casey M. Jelsema, Research Fellow, Biostatistics Branch, NIEHS):

In many applications, such as in dose-response studies or time-course experiments, researchers are interested in testing for specific inequality constraints or patterns among the means of experimental groups. This R package is designed to test for such inequality patterns using a robust residual bootstrap based methodology which does not require the data to be normally distributed. Furthermore, this software can also handle the situation when covariates and/or random effects are present. Thus, for example, this package can be used in the context of repeated measurement designs with covariates. This package comes with a user friendly graphical interface so no programming is necessary to run this package. All the user needs to do is to provide input source of the data and select options from the interface.




• Jelsema C, Peddada SD. An R Package for Linear Mixed Effects Models under Inequality Constraints. Journal of Statistical Software doi: 10.18637/jss.v075.i01 (2016). 



Shyamal D. Peddada