Genes that were highly connected to many of the trait measures were entered into gene set over-representation analysis using WebGestalt system which performs a hyper-geometric test for the enrichment of Gene Ontology and KEGG pathway annotations among sets of genes. Among the top most highly connected 5% of genes, significant over-representation (P-value < 1E–05) was observed in 30 pathways of relevance to brain and behavior. The top most highly connected 10% of genes showed significant over-representation (P-value < 1E–05) of 32 pathways related to brain and behavior (Table 4). These results indicate the pathways which are either candidate causes or under regulation of genetic polymorphisms that are commonly observed in the BXD population. The principle behind this approach is that genetic variation affects many genes associated with QTL candidates, and that the convergent evidence gained from multi-dimensional analyses of the same traits can highlight the most frequently associated pathways. Because these data have all been deposited in the ODE tool, users of these data can perform analyses of specific and roughly balanced subsets of the traits.
