Masters candidate in Biostatistics, Karan Shah, will present:

“Estimating SNP Heritability of Proteomic Biomarkers”

Plan B Adviser: Cavan Reilly

Abstract: Heritability is a measure of how well the variance in various traits can be accounted for by genes. Genome-based restricted maximum likelihood (GREML) is a method for estimating the phenotypic variance explained by all SNPs in unrelated individuals. Here, GREML, as implemented in the GCTA software is used to estimate the heritability of various proteomic biomarkers. We have data from genotypic and phenotypic data on 1,493 individuals for approximately 450 biomarkers across 5 OLINK proteomic panels. In order to account for population structure, the PC-RELATE method was used in order to construct the genetic-relatedness matrix (GRM) and the estimates obtained were compared with the GRM constructed using GCTA. Different relatedness cutoff thresholds were also used to assess the sensitivity of estimates. It was established that the results are highly sensitive to previously mentioned analysis decisions and thus, care must be taken while interpreting the obtained heritability estimates.

Presented via Zoom: https://umn.zoom.us/j/94769488267
Meeting ID: 947 6948 8267