The Department of

Statistics and Operations Research

The University of North Carolina at Chapel Hill

Graduate Student Seminar

Friday, April 14th, 2023 120 Hanes Hall 3:30-4:30pm

or via Zoom:

https://unc.zoom.us/j/92731259806?pwd=eTc3Ylo0SWNFOHV4eERvSjRrUmlKQT09

Meeting ID: 927 3125 9806

Passcode: 100210

Vasileios Katsianos

University of Chicago – Department of Statistics

Association Analysis of Phenotypes Jointly Influenced by a Pair of Organisms

Abstract: The virulence of an infectious disease is usually affected by a combination of a host and at least one pathogen organism, so it is influenced by all of their genomes. Previous experiments have divulged that combining genetic information from all organisms involved has enabled to identify more relevant genetic variants than just individually performing an association analysis on each organism. Conducting genetic mapping on a pathogen organism usually presents some challenges of its own. Random insertions and deletions in the pathogen genome can lead to the existence of multiallelic genetic variants, which are often associated with the presence of some alleles with particularly small frequencies. We build upon existing work to construct a pathogen genetic relatedness matrix which aggregates information across genetic variants with different numbers of alleles and accommodates small allele frequencies without introducing numerical instability by applying some form of regularization. When considering interaction effects between the genotypes of different organisms on the phenotype, the fact that the phenotype is held fixed while the two genotypes vary across different interaction tests may often lead to significant departures from the known null distribution of the test statistic. We present diagnostic tools to detect such departures and a method to correct for them under a wide range of scenarios. We demonstrate the applicability of our methodology on data from a hepatitis C clinical trial and simulated data.

There will be happy hour following this talk.

Share this Event