Excerpt
Establishing the pathogenicity of variants found in clinical whole-exome sequencing is challenging. Functional testing is a relatively slow process that often involves scientific collaborations with specialized laboratories. In order to avoid this process, Majithia et al. embarked on a seemingly heroic effort to prospectively characterize the functional consequences of each possible missense variant in the
PPARG gene (Majithia et al.
2016). Autosomal dominant mutations in
PPARG cause familial partial lipodystrophy, a disorder characterized by abnormal fat distribution and type 2 diabetes. Common variants in
PPARG have also been associated with obesity and type 2 diabetes in genome-wide association studies (GWAS). Indeed,
PPARG plays a crucial role in fat development and was coined as the ultimate thrifty gene (Auwerx
1999). Using highly parallel oligonucleotide synthesis, Majithia et al. first constructed a complementary DNA (cDNA) library containing 99% of all possible amino acid substitutions in the peroxisome proliferator activated receptor gamma (PPARG) protein (Majithia et al.
2016). This library was then introduced in a human macrophage cell line that was rendered PPARG deficient through genome editing. Activity of the PPARG variant proteins could be conveniently assessed by fluorescence-activated cell sorter (FACS), allowing the cells to be sorted in high- or low-activity PPARG bins. PPARG variant distribution in each bin was subsequently determined by sequencing and then used to generate a function score for each variant. Next, a classifier to predict the likelihood of a new variant being benign or pathogenic was established using data on the pathogenicity of previously characterized variants. Importantly, the authors show that this classifier annotated almost all incidentally identified variants as benign and was able to accurately predict benign and pathogenic variants in a cohort of patients with partial lipodystrophy (Majithia et al.
2016). The diagnosis of inborn errors of metabolism (IEM) could also benefit from prospective validation of variants by directly assessing gene or pathway function. First, mutations causing IEM are often private and thus have not been functionally assessed. Second, metabolite perturbations are not always the best predictor of gene or pathway function. Finally, functional analyses are often performed across multiple laboratories using various techniques. The consistent evaluation of all variants in a single assay will make comparison of individual variants more reliable. …
