We applied conditional analysis via the FUSION framework (FUSION.post_process.R) to the regions of significant genes after correction to investigate the proportion of implicated genes in any given locus that are independently associated. Specifically, jointly significant genes retain their significance after jointly estimating association for all models within a 500 000 base pair region of a significant gene. Marginally associated genes, which are not jointly significant, likely arise due to factors such as genes for which predicted expression was correlated.
