Genome-wide association studies is a statistic technology to detect association between a genetic variant and a trait value (e.g., disease status). For one genetic variant (independent variable x) and one response variable (y), PROC GLIMMIX can do the job needed. But we often have 10 thousands or millions of genetic variants. Lopping over PROC GLIMMIX for all variants is extremely slow. Can we develop a new SAS procedure to scan the entire genome faster than using %macro looping over PRCO GLIMMIX?
