There is also a fundamental difference in the way that GWAS and HBCGM results are interpreted. The results from a conventional murine GWAS are usually evaluated without considering other types of data, and a very small genome-wide significance cutoff is applied in order to strictly control the false positive rate. Therefore, a large sample size is required to reduce the number of SNPs that will randomly correlate with a trait value to enable a true causative locus to be identified. However, the sample size used in prior murine GWAS studies was always less than 20 strains (and often 6–10 strains are used), which makes it difficult for even a true causative locus to have a p-value that achieves genome-wide significance level. Thus, the need to control the false positive rate leads to a high probability that the study will produce a false negative result, which explains why investigators have concluded that murine association studies can’t work. In contrast, HBCGM results are viewed within the context of integrated analysis of a biomedical trait. As a consequence, a less stringent filtering criterion
