Contrasting Genetic Architectures using Fast Variance Components Analysis

Posted on Feb 07, 2024

This note is for Loh, P.-R., Bhatia, G., Gusev, A., Finucane, H. K., Bulik-Sullivan, B. K., Pollack, S. J., de Candia, T. R., Lee, S. H., Wray, N. R., Kendler, K. S., O’Donovan, M. C., Neale, B. M., Patterson, N., & Price, A. L. (2015). Contrasting genetic architectures of schizophrenia and other complex diseases using fast variance components analysis. Nature Genetics, 47(12), 1385–1392.

Heritability analyses of GWAS cohorts have yielded important insights into complex disease architecture, and incresing sample sizes hold the promise of further discoveries.

analyze the genetic architecture of schizophrenia in 49806 samples from the PGC, and nine complex diseases in 54734 samples from the GERA cohort.

develop a fast algorithm for multi-component, multi-trait variance components analysis that overcomes prior computational barriers

variance components analysis has had considerable impact on research in human complex trait genetics

BOLT-REML, apply it to analyze ~50,000 samples

BOLT-REML algorithm

• $O(MN^{1.5})$ time
• existing stanard methods: $O(MN^2+N^3)$ time

for memory

• BOLT-REML only requires MN/4 bytes of memory (nearly independent of the number of variance components used)
• standard REML analysis requires $O(N^2)$ memory per variance component