In order to link PPAR agonist-regulated genes with changes in alcohol drinking, we tested the differentially expressed genes for overlap with a list of genes known to affect alcohol consumption (based on studies of mutant mice). In the amygdala, feno and tesa, but not beza treatment groups, contained more alcohol-related genes than expected by chance (hypergeometric p < 0.007 for feno and tesa; p = 0.46 for beza). We used the non-effective beza treatment to filter out non-related genes and found that genes regulated by both feno and tesa but not beza had a greater number of alcohol-related genes than expected by chance in both the amygdala and PFC (Figure 3). These genes have already been experimentally validated and shown to change alcohol consumption, thus supporting our approach of using the non-effective treatment to filter out non-related genes.
