Jeffrey Townsend, PhD

Associate Professor of Public Health (Biostatistics) and of Ecology and Evolutionary Biology; Director of Bioinformatics, Yale Center for Analytical Sciences

Departments & Organizations

Cancer Center, Yale: Genomics, Genetics, and Epigenetics

Office of Student Research

School of Public Health: Biostatistics

Yale Combined Program in the Biological and Biomedical Sciences (BBS): Computational Biology and Bioinformatics: Computational Approaches to Functional and Integrative Genomics | Microbiology: Molecular Genetics


Professor Townsend received his Ph.D. in 2002 in organismic and evolutionary biology from Harvard University, under the advisement of Daniel Hartl. His Ph.D. was entitled "Population genetic variation in genome-wide gene expression: modeling, measurement, and analysis", and constituted the first population genetic analysis of genome-wide gene expression variation. After making use of the model budding yeast S. cerevisiae for his Ph.D. research, Dr. Townsend accepted an appointment as a Miller Fellow at the University of California-Berkeley in the Department of Plant and Microbial Biology, where he worked to develop molecular tools, techniques, and analysis methodologies for functional genomics studies with the filamentous fungal model species Neurospora crassa, co-advised by Berkeley fungal evolutionary biologist John Taylor and molecular mycologist Louise Glass. In 2004, he accepted his first appointment as an Assistant Professor in the Department of Molecular and Cell Biology at the University of Connecticut. In 2006 he was appointed as an Assistant Professor the Department of Ecology and Evolutionary Biology at Yale University, and in 2013 he was appointed as an Associate Professor of Biostatistics in the Yale School of Public Health.

Education & Training

PhD Harvard University (2002)
ScB Brown University (1994)
Miller Postdoctoral Fellw University of California, Berkeley

Honors & Recognition

  • Young Investigator's PrizeAmerican Society of Naturalists (2005)

  • Ph.D. Thesis ranked among the Top Four Life Science ThesesCouncil of Graduate Schools (2002)

  • Walter M. Fitch Prize for Best Young InvestigatorSociety for Molecular Biology and Evolution (2001)

Professional Service

  • Faculty Advisor Yale College Road Runners (2008)

International Activity

  • Ebolavirus sequence analysis and epidemic modeling Freetown, Sierra Leone (2014 - 2015)

    Using Ebolavirus genomic and epidemiological data, we are conducting joint analyses of both data types to fit dynamic transmission models for the major Ebolavirus outbreak in Sierra Leone. Our analysis evaluates the degree of underreporting of disease cases, and the degree of social clustering of...

  • Yaws Brazzaville, Republic of the Congo (2012)

    Dynamic mathematical modeling predicting treatment necessary for eradication of Yaws

  • Human Hookworm Zanzibar, Tanzania (2011)

    Dynamic mathematical modeling evaluating the efficacy of treatment of Human Hookworm Infection

  • Antibiotic resistance Tromso, Norway (2010 - 2014)

    Evaluate trace antibiotics in the environment as an environmental toxin

Edit this profile

Contact Info

Jeffrey Townsend, PhD
Lab Location
Laboratory of Epidemiology and Public Health
60 College Street, Ste 7th Floor

New Haven, CT 06510
View on map...
Office Location
135 College Street, Fl Floor 2 Ste Suite 200 Rm Room 222
New Haven, CT 06510
View on map...
Mailing Address
135 College St, Room 222
New Haven, CT 06510-2483

Townsend Lab

Research Image 1

Spotted DNA microarray. Red spots represent genes abundantly expressed in wine yeast growing in a high concentration of copper sulfate. Green spots represent genes expressed abundantly in wine yeast growing at a low concentration of copper sulfate. Copper sulfate is often applied in vineyards to control growth of fungi.

Research Image 2

Information on the Tree of Life. Recent efforts to reveal the evolutionary history of life on earth have increasingly relied on the sequencing of DNA from multiple species for multiple genes. This figure demonstrates a principle that should guide these efforts: to understand deep divergences, sample taxa that diverge deeply first. a) and b) Curves depict the cumulative support for the bold deep internode of four species (the fungi Yarrowia lipolytica, Saccharomyces cerevisiae, Coccidioides immitis, and Neurospora crassa), ranging from zero to complete sampling for several sampling schemes: the outcome based on perfect and worst-possible performance (dashed); outcome based on prioritizing sampling based on an novel theoretical prediction using rate of evolution of the sequences (solid); outcome based on prioritizing sampling of all genes for the deepest ingroup (dash-dotted); expectation for haphazard sampling (dotted). c) The established chronogram, or time tree, of the evolution of these species. Vertical bars in the plots correspond to switches from sampling characters from deeper-branching to sampling characters from shallower-branching taxa; note that the slope of the increase in cumulative information (red and green curves) declines as sequences are sampled from more recently diverged lineages in the tree, and that this pattern of high utility to sampling the deepest lineages is revealed for both the clade in panel a and the clade in panel b.

Research Image 3

Population genetic modeling of HGT suggests several key quantities are important to designing any sampling-based assay of horizontal gene transfer (HGT) in large populations. The HGT rate r and the exposed fraction X play significant but ultimately minor roles in the population dynamics, most likely impacting only the number of original opportunities for horizontal spread of genetic material. The malthusian selection coefficient m of the transferred genetic material and the time in recipient generations t from exposure play key, non-linear roles in determining the potential for detection of HGT. Sample size n is important, but frequently the practical sample sizes to be obtained are many orders of magnitude below the extant population size. It is therefore essential to wait until natural selection has had time to operate, so that it is essential to wait until natural selection has a chance to operate to have any chance of effectively detecting horizontal gene transfer events.