Assuming an ideal set of circumstances, a quantitative statistical assessment could be performed to determine whether the 51 iMO–human-AUD gene set is larger than would be expected by chance. For example, given 293 iMO–human genes and 732 human genes in the query sample, and assuming 21,000 total human genes (Harrow et al., 2012), one would expect 293/21,000 × 732 ≈ 10 genes in the iMO–human-AUD set by random chance alone. As the observed iMO–human-AUD set contains 51 genes, this would correspond to an approximately 5-fold overrepresentation. Unfortunately, analyses of this type are not strictly valid for several reasons. First, the sets of 293 iMO–human and 732 human alcoholism-related genes queried for overlap are not independent, which is evidenced by published connections between studies in iMOs and humans for several individual genes (the ADH family, AUTS2, ALK, GPC5, GABBR1, NPY, etc.). Second, not all of the genes appearing in the HuGe Navigator have been associated with alcoholism, alcohol dependence, or other forms of AUD. Specifically, we found that of 83 genes in the HuGe Navigator evaluated as part of this review,
