Quantification of Phenotype Information Aids the Identification of Novel Disease Genes.

Next-generation sequencing led to the identification of many potential novel disease genes. The presence of mutations in the same gene in multiple unrelated patients is, however, a priori insufficient to establish that these genes are truly involved in the respective disease. Here, we show how phenotype information can be incorporated within statistical approaches to provide additional evidence for the causality of mutations. We developed a broadly applicable statistical model that integrates gene-specific mutation rates, cohort size, mutation type, and phenotype frequency information to assess the chance of identifying de novo mutations affecting the same gene in multiple patients with shared phenotype features. We demonstrate our approach based on the frequency of phenotype features present in a unique cohort of 6,149 patients with intellectual disability. We show that our combined approach can decrease the number of patients required to identify novel disease genes, especially for patients with combinations of rare phenotypes. In conclusion, we show how integrating genotype-phenotype information can aid significantly in the interpretation of de novo mutations in potential novel disease genes.

Duplications of SLC1A3: Associated with ADHD and autism.

We report four patients with a similar gain in 5p13.2 encompassing a single gene: SLC1A3. Behavioural problems resembling ADHD and/or autism-like features are observed which is in line with the glial glutamate transporter role of SLC1A3. We consider an association between SLC1A3 and the behavioural problems which can also be considered a contributing factor to behavioural problems in larger duplications overlapping the 5p13 microduplication syndrome region.