Welcome to the Computational Genetics Laboratory in the Institute for Biomedical Informatics at The Perelman School of Medicine of The University of Pennsylvania in Philadelphia, PA, USA.

We are affiliated with the Division of Informatics in the Department of Biostatistics and Epidemiology, the Department of Genetics, and the University of Pennsylvania Health System.

We are also part of the graduate training programs in Genomics and Computational Biology (GCB) and Epidemiology and Biostatistics.

For our latest news see Epistasis Blog or follow along on Twitter.



Our research goal is to develop, evaluate and apply novel computational methods and open-source software for identifying genetic and genomic biomarkers associated with human health and disease. Our focus is on methods that embrace, rather than ignore, the complexity of the genotype-to-phenotype mapping relationship due to phenomena such as epistasis and plastic reaction norms. Areas of interest include artificial intelligence, bioinformatics, biomedical informatics, complex systems, computational biology, genetic epidemiology, genomics, human genetics, machine learning, and visual analytics.

Our education goal is to provide interdisciplinary training and research experience to undergraduate, graduate, and postgraduate students. Our philosophy is that biomedical researchers of the future need to speak multiple languages to effectively collaborate with diverse teams of people focused on solving the hardest problems in health and healthcare.


A central focus of our research is the detection, characterization and interpretation of epistasis or gene-gene interaction. The word epistasis was coined by William Bateson in the early 1900s in his book Mendel's Principles of Heredity and was used to describe deviations from mendelian inheritance patterns due to one gene standing upon or modifying the effects of another gene.

Our working hypothesis is that epistasis is a ubiquitous component of the genetic architecture of human health and disease due to the importance of biomolecular interactions that drive biological systems in a genotype-dependent manner. Canalization may provide an evolutionary explanation for why epistasis is so common.

We make the distinction between biological and statistical epistasis. Biological epistasis is what Bateson had in mind and is the result of biomolecular interactions at the cellular level. In contrast, statistical epistasis is a population-level summary and is more similar to Sir Ronald Fisher's definition of epistacy or deviation from additivity in a linear statistical model. Our goal is to develop and use data mining and machine learning algorithms to model statistical epistasis. The ultimate goal is to infer biological epistasis using systems biology approaches such as network modeling.