Utility of exome-based gene panel testing and research analysis of existing exome sequencing data in suspected monogenic diabetes

Background Monogenic diabetes (MD) accounts for 1–4% of all diabetes cases in children and adolescents; however, the diagnostic rate of patients with MD remains low. Identification of molecular causes of MD is important to ensure appropriate patient treatment and genetic counseling. In this study, we sought to investigate the diagnostic yield of five MD-related multigene panels in 83 patients who previously received clinical exome-based gene panel (EBGP) testing. In addition, we conducted research analysis of previously generated exome sequencing (ES) data to determine whether it may increase diagnostic yield and/or lead to the discovery of biologically suspicious variants in genes not yet associated with MD. Methods Research exome analysis was conducted for patients that previously received clinical EBGP testing. Variants of interest were categorized as follows: (1) detection of a variant associated with indication for testing, but not in a gene included on the ordered clinical panel (candidate diagnostic finding); (2) variants in candidate diabetes disease genes; and (3) clinically actionable incidental findings. Patient clinical information was taken from electronic health records. Results Clinical genetic testing data representing 83 endocrine patients (81% pediatric, 55% female) suspected of having MD was analyzed. Initial EBGP clinical testing yielded the following results: Seven patients (8%) were positive, 37 (45%) had inconclusive results due to variants of uncertain significance, and 39 (47%) were negative. The most frequently observed diagnostic EBGP results included WFS1 pathogenic variants (2/7 patients) and GCK pathogenic variants (2/7 patients). Analysis of previously generated exome sequencing analysis identified 12 variants of interest across 10 patients. One patient was found to harbor a pathogenic variant in IGF1R , which was outside the ordered gene panel and considered a candidate diagnostic finding given the patient’s short stature. Another patient harbored compound heterozygous pathogenic variants in ARSA , considered a clinically actionable incidental finding at the time of detection. Nine variants were found in genes (including CAPN10 and SLC2A4 ) not associated with diabetes but with evidence supporting involvement in diabetes-related pathways (evidence includes animal models, computational predictions, pathway enrichment, gene network modeling, and functional annotation predictions). Conclusion In our cohort of suspected MD, research analysis of existing ES data resulted in a diagnostic yield comparable to current EBGP testing (i.e., no additional diagnostic variants were detected). However, this study provided the added benefit of detecting variants in candidate disease genes that might eventually lead to the discovery of new diabetes-associated genes.