Statistics Seminar
| Date/Time: | Friday, 14 Nov 2014 from 11:00 am to 11:50 am |
|---|---|
| Location: | Gilman 2109 |
| Cost: | Free |
| URL: | www.stat.iastate.edu |
| Contact: | Jeanette La Grange |
| Phone: | 515-294-3440 |
| Channel: | College of Liberal Arts and Sciences |
| Categories: | Lectures |
| Actions: | Download iCal/vCal | Email Reminder |
The tidal wave of next-generation sequencing (NGS) data has arrived, but more questions than answers exist about how to best analyze NGS data to investigate the potential contribution of rare genetic variants to human disease. Numerous rare variant association testing methods have been proposed which attempt to aggregate association signals across multiple variant sites in an effort to increase statistical power. While emerging simulation results suggest that some rare variant testing methods work better than others for particular genetic architectures, little concrete understanding of the tests is available. We recently proposed a geometric framework which quickly classifies existing rare variant tests of association into two broad categories: length and joint tests. We then demonstrated how genetic architecture (relative risk distribution, allele frequency distribution and number of variants) directly relates to the behavior of length and joint tests. In this talk I will introduce the geometric framework and then illustrate how it can be used to predict test behavior, motivate alternative rare variant testing strategies, evaluate the impact of genotyping error and give insights into post-hoc approaches.