November Webinar: Chirag Patel, “Building a search engine to identify environmental exposures in disease and phenotypes”

Sep 29, 2017 | PROTECT Team, PROTECT Webinar Series

On Monday, November 6th, PROTECT hosted the first webinar of the 2017-2018 academic schedule. This webinar featured Chirag Patel, PhD, Assistant Professor of Biomedical Informatics at Harvard Medical School, who spoke on “Building a search engine to identify environmental exposures in disease and phenotypes.”

Read on to learn more about the webinar and Dr. Chirag’s impressive bio, or listen to the recording here.


In an anticipated new era of high-throughput environmental exposure assessment, there is an urgent need for new analytics approaches to drive discovery of new exposures associated with disease and phenotype. However, we lack analytic tools and data to discover new exposures to explain missing phenotypic variation in the population. In contrast, genomic investigations have accelerated the pace for discovery of inherited factors in disease at rapid pace and the same advances should be applied to discover exposures in disease.

In this talk, we present new analytic approaches to drive discovery of new exposures in disease and the missing phenotypic variation in the population. Our proposed approaches offers a few advantages, including explicitly mitigating false positive findings and assessing the entire database of potential environmental correlates systematically to avoid a fragmented literature of associations. However, there are many analytic and biological challenges to be addressed. Most studies are observational studies and prone to biases such as confounding and reverse causality (e.g., disease coming before the exposure). Yet another issue includes multiple significant associations (low pvalues) with small effect sizes. This may be the norm in exposome-wide analytics, where correlations may be small but seemingly correlated with everything else and it will be a challenge to ascertain what causes what. Despite these challenges, a new big data paradigm is required to find new causative factors for disease risk.


Chirag Patel’s long-term research goal is to address problems in human health and disease by developing computational and bioinformatics methods to reason over high-throughput information spanning molecules to populations. Chirag’s group focuses on computational strategies to efficiently and reproducibly uncover the complex interaction between the exposome, genome, and phenometowards development of new tools for disease diagnosis and therapy. He trained in biomedical informatics at Stanford University. Prior to graduate work, Chirag was a software engineer in the biotechnology industry. Chirag is now an assistant professor of biomedical informatics at Harvard Medical School where he teaches introductory courses in data science and is a mentor to 3 post-doctoral associates and advises 3 PhD students. Follow him on twitter at @chiragjp or visit his website:

Attendees at the Northeastern University viewing of Patel’s webinar for the PROTECT Center