Learning Reduced Representations of Genomic Data

Seminar Date: 
Seminar Time: 
11am - 12pm
Seminar Location: 
5607 Baum Boulevard, Room 407A
Maria Chikina, PhD
Presenter's Institution: 
University of Pittsburgh

Genome scale technologies can measure thousands or even millions of molecular features but the individual measurements are often noisy redouts of a smaller set of parameters describing the internal biological state.  Indeed, many computational analyses can be viewed as methods to infer the true biological variables from their coordinated effects on a larger set of experimental readouts.  In this talk I will discuss methods that use biologically motivated priors and constraints to  reduce large scale measurements to a smaller set of interpretable features facilitating both downstream computation and mechanistic reasoning.