Accounting for genetic heterogeneity in homozygosity mapping: application to Mendelian susceptibility to mycobacterial disease.

INTRODUCTION: Genome-wide homozygosity mapping is a powerful method for locating rare recessive Mendelian mutations. However, statistical power decreases dramatically in the presence of genetic heterogeneity.
METHODS: The authors applied an empirical approach to test for linkage accounting for genetic heterogeneity by calculating the sum of positive per-family multipoint LOD scores (S) across all positions, and obtaining corresponding empirical p values (EmpP) through permutations.
RESULTS: The statistical power of the approach was found to be consistently higher than the classical heterogeneity LOD by simulations. Among 21 first-cousin matings with a single affected child, for five families linked to a locus of interest and 16 families to other loci, S/EmpP achieved a power of 40% versus 28% for heterogeneity LOD at an α level of 0.001. The mean size of peak linkage regions was markedly higher for true loci than false positive regions. The S/EmpP approach was applied to a sample of 17 consanguineous families with Mendelian susceptibility to mycobacterial disease, leading to the identification of two mutations in IL12RB1 and TYK2 from the largest of six linkage regions at p<10(-3).
CONCLUSIONS: The S/EmpP approach is a flexible and powerful approach that can be applied to linkage analysis of families with suspected Mendelian disorders.