Our method which explicitly accounts for important confounders on the association scores of genes (e.g. gene size) and gene sets (e.g. positional effects of genes in a gene set) in the absence of genotype data, and that provides upper-bound estimates of number of associations per gene set, should provide accurate tests of gene sets of interest, especially for analyzing large GWA scan meta-analyses. MAGENTA can also be applied to sets of genetic loci other than genes, such as linkage disequilibrium blocks. More generally, such GSEA approaches may be valuable for gene and pathway analysis of other types of genetic studies that deal with multiple measurements per gene, such as exon resequencing in case-control studies.
