A Low-Rank representation and Sparse regression for eQTL mapping. This algorithm accounts for confounding factors such as unobserved covariates, experimental artifacts, and unknown environmental perturbations.
doi: 10.1093/bioinformatics/btt075
please use demo to get start.
Suppose you have expression data Y and SNP data X You can call function LORSmain as
[B, L, mu] = LORSmain(Y,SNP)
It will return you the estimated coefficient matrix B, the low-rank matrix L and intercept mu.