If depression is driven by many thousands of loci of weak effect, another strategy may be to combine genetic signals across many SNPs into functionally-defined gene sets or pathways. Pathway approaches can be considerably more powerful than single variant analyses, as the aggregation of weak signals from multiple causal variants may yield statistically significant evidence in support of a given gene or pathway.111,112 Thus far, investigators have primarily examined pathways related to specific biological functions (e.g., axon guidance, cell functioning) as defined by human-curated bioinformatics resources, such as the Kyoto Encyclopedia of Genes and Genomes (KEGG)113 or Gene Ontology.114 Recent studies of candidate gene pathways have found evidence that genes involved in glutamatergic synaptic neurotransmission,115 among others,116 were significantly associated with depression. Evidence in support of gene sets or pathways also comes from several GWAS described previously and shown in Table 2, which found significant support for some pathways.46,56 One of the major drawbacks of gene-set analyses is that they require predefined sets of genes. Gene sets defined by current annotation databases, such as KEGG or GO, vary in their
