Using this approach, however, we did not find any association that was significant at the genome-wide Bonferroni-corrected threshold of 1.8 × 10−7 (Fig. 2A; Supplementary Material, Table S2). Nevertheless, some of the most strongly associated SNPs are within or near genes that are involved in cellular proliferation or lymphocyte function. For example, the strongest signal [P-value = 4.32 × 10−7; false discovery rate (FDR) = 0.12] is at rs2746347 (Fig. 2B; Supplementary Material, Table S2), a SNP located in the first intron of PRKAA2 on chromosome 1. The gene product of PRKAA2 is the catalytic subunit of the adenosine monophosphate-activated protein kinase, an enzyme that regulates the cell cycle in response to glucose availability (28), which is up-regulated upon T cell receptor stimulation (29). The second strongest signal is at rs881827 (Fig. 2C; Supplementary Material, Table S2), located within an intron of S100B. This gene is thought to play a role in inflammatory responses and is secreted by activated T lymphocytes (30). These observations suggest that our GWAS approach is underpowered to identify loci with key roles in determining variation
