Description
Mathematics Colloquium.
Speaker: Nilanjan Chatterjee.
Affiliation: Johns Hopkins University.
Date and Time: Wednesday 26 February, 04:00 pm - 05:00 pm.
Venue: Ramanujan Hall, Department of Mathematics.
Title: Disease risk prediction and causal inference using genome-wide
genetic data.
Abstract: Recent genome-wide association studies have led to
identification of thousands of genetic variants associated with complex
traits and diseases like adult height, body mass index, heart disease,
type-2 diabetes and cancer. The large scale genetic data, some of which
are publicly available, provide statisticians, mathematicians, computer
scientists and other quantitative researchers an incredible opportunity
for the development and applications of novel methods and algorithms. In
this talk, I will describe the work from our laboratory to harness the
power of these big datasets to address two most pressing problems in
public health research. In particular, I will describe simple and more
advanced machine learning methods for building genetic risk-scores from
these datasets that can be used to predict prospectively individuals' risk
of diseases. Further, I will describe how genetic data can be used to
conduct, "instrumental" variable analysis, an approach popular in
Economics, to understand causal relationship among risk-factors and health
outcomes.