A Global Test of Association between Non-Hodgkin Lymphoma and SNPs in Histone-pathway Genes

We analyze data from a population-based case-control study to test whether NHL is associated with SNPs in histone-pathway genes. When individual SNP associations are minor, the standard approach of testing SNPs one-at-a-time and then correcting for multiple testing has low power. Our global testing approach avoids the multiple-testing penalty by modelling random SNP effects and testing variance. We show how an approximate score statistic may be derived by writing the likelihood as an expected conditional likelihood given latent genetic values for each individual and then applying a Taylor-series approximation. The resulting statistic is applied to the NHL data and its statistical significance is evaluated using a permutation-based procedure. Our results add to growing evidence that the histone pathway plays a role in NHL.