SEIR: a novel multi-locus GWAS method that provides higher statistical power for fast identifying variant-phenotype associations

Abstract

AbstractMulti-locus genome-wide association study (GWAS) methods have considered the joint effects of multiple variants to more accurately unravel the genetic basis of complex traits. Here, we developed a novel multi-locus GWAS method named Selector-Embedded Iterative Regression (SEIR), which integrates the embedded selector with fast single-marker scanning in an iterative manner. SEIR has excellent adaptability and flexibility under various genetic architectures for qualitative and quantitative traits. Reliability of SEIR was experimentally supported by integrating GWAS with 3D epigenomics in a real trait. Conclusively, SEIR exhibits higher statistical power for fast identifying putative variants compared to other single- and multi-locus methods.