Dalla Lana School of Public Health

The Biostatistics Seminar Series presents:

“Cost-reduction Strategies for Post-GWAS Next-generation Sequencing Studies” by Dr. Osvaldo Espin-Garcia, Princess Margaret Cancer Centre

Abstract: Genome-wide association studies (GWAS) have greatly helped in advancing our understanding of the genetic underpinnings of human health and disease by identifying thousands of genetic loci associated with a multitude of traits. Although powerful at pinpointing genomic regions associated with traits, GWAS typically fail to identify causal loci, as such, researchers increasingly turn to next-generation sequencing (NGS) technologies to complement GWAS data. However, despite the plummeting cost of NGS in recent years, the cumulative cost of NGS data accrual remains prohibitive for many projects. The two-phase study provides a cost-reduction strategy for NGS data collection by selecting an informative subsample based on already available (and often inexpensive) phase 1 data, e.g. outcomes, GWAS. Then, NGS data are collected on the informative subsample alone, reducing the overall cost. Finally, inference is performed via missing data methods using information available from phases 1 and 2. In this talk I will cover recent and ongoing work on two-phase study design and analysis for NGS data accrual. First, I will introduce a genetic algorithm to select an informative subsample under a sample size constraint for single region post-GWAS analysis. Second, I will show results on the use of polygenic risk scores to guide two-phase design and analysis in multi-regional settings. These approaches will be illustrated using the Northern Finland Birth Cohort of 1966. Extensions to multivariate outcomes and other ‘omics technologies will be discussed

For Dr. Espin-Garcia’s biosketch, please see https://www.dlsph.utoronto.ca/faculty-profile/osveg/

Register in advance for this seminar via
https://phesc.zoom.us/meeting/register/tZAkfuqsrTgqGNV_oKEHAM7rqR6YjdOJyptQ