Incorporating prior biological information in linkage studies increases power and limits multiple testing

We used the Genetic Analysis Workshop 15 Problem 1 data set to search for expression phenotype quantitative trait loci in a highly selected group of genes with a supposedly correlated role in the development of the enteric nervous system. Our strategy was to reduce the level of multiple testing by analyzing at the genome-wide level a limited number of genes considered to be the most promising enteric nervous system candidates on the basis of mouse expression data, and then extend the analysis to a larger number of traits only for a small number of candidate linked regions. Such a study design allowed us to identify a "master regulator" locus for several genes involved in the enteric nervous system, located in 9q31. In particular, one of four traits included in the genome-wide analysis and 2 of 57 from the follow-up single-chromosome analysis showed LOD scores above 2 around position 109 on chromosome 9 by univariate variance-component linkage analysis. Bivariate linkage analysis further supported the presence of a common regulatory locus, with a maximum multipoint LOD score of 5.17 and five additional LOD scores > 3 in the same region. This region is particularly interesting because a susceptibility locus for Hirschsprung disease, a disease characterized by enteric malformation, was previously mapped to 9q31. The proposed strategy of limiting the genome-wide analysis to a small number of well characterized candidate expression phenotypes and following up the most promising results in a larger number of correlated traits may prove successful for other groups of genes involved in a common pathway.