Remarkably, all six genes originally selected from large and diverse gene expression datasets were associated with changes in alcohol consumption in mice. Apart from the gene expression studies, none of these genes had previously been implicated in alcohol actions and most are ‘immune/inflammatory’ genes with limited visibility in neurobiology. These findings support the validity of our strategy for selection of gene targets from microarray data using convergent statistical, informatics and functional approaches. A unique aspect of the present study is the behavioral validation for use of multiple microarray datasets to predict targets important for alcohol action and indeed the actual validation of this approach for any behavioral trait. It is important to note that the most consistent and robust changes occurred in the 24 hr two-bottle choice test, the test that was the basis for selection of these genes. The 24 hr two-bottle choice test showed reduced alcohol preference and consumption in most male and female null mutant mice. However, in limited access 1B-DID tests, only deletion of the Il1rn and Ctss genes decreased ethanol consumption, indicating differences in the genetic control of these behaviors (Supplemental Table 10).
