Symptom-driven idiopathic disease gene identification.

Abstract

Rare genetic variants are the major cause of Mendelian disorders, yet only half of described genetic diseases have been causally linked to a gene. In addition, the total number of rare genetic diseases is projected to be far greater than that of those already described. Whole-genome sequencing of patients with subsequent genetic and functional analysis is a powerful way to describe these gene anomalies. However, this approach results in tens to hundreds of candidate disease-causative genes, and the identification of additional individuals suffering from the same disorder can be difficult because of rarity and phenotypic heterogeneity. We describe a genetic network-based method to rank candidate genes identified in family-based sequencing studies, termed phenotype informed network (PIN) ranking. Furthermore, we present a case study as an extension of the PIN ranking method in which disease symptoms drive the network ranking and identification of the disease-causative gene. We demonstrate, through simulation, that our method is capable of identifying the correct disease-causative gene in a majority of cases. PIN-rank is available at https://genomics.scripps.edu/pinrank/. We have developed a method to prioritize candidate disease-causative genes based on symptoms that would be useful for both the prioritization of candidates and the identification of additional subjects.