Statistical evidence for high‐penetrance MODY‐causing genes in a large population‐based cohort

AIMS: Numerous genes have been proposed as causal for maturity‐onset diabetes of the young (MODY). Scoring systems to annotate mutation pathogenicity have been widely used; however, statistical evidence for being a highly penetrant MODY gene has not been well‐established. METHODS: Participants were from the UK Biobank with whole‐exome sequencing data, including 14,622 with and 185,509 without diagnosis of diabetes. Pathogenic/likely pathogenic (P/LP) mutations in 14 reported and 3 possible MODY genes were annotated using American College of Medical Genetics criteria. Evidence for being a high‐penetrant MODY gene used two statistical criteria: frequency of aggregate P/LP mutations in each gene are (1) significantly more common in participants with a diagnosis of diabetes than without using the SKAT‐O (p < .05) and (2) lower than the maximum credible frequency in the general population. RESULTS: Among the 17 genes, 6 (GCK, HNF1A, HNF4A, NEUROD1, KCNJ11 and HNF1B) met both criteria, 7 (ABCC8, KLF11, RFX6, PCBD1, WFS1, INS and PDX1) met only one criterion, and the remaining 4 (CEL, BLK, APPL1 and PAX4) failed both criteria, and were classified as ‘consistent’, ‘inconclusive’ and ‘inconsistent’ for being highly penetrant diabetes genes, respectively. Diabetes participants with mutations in the ‘consistent’ genes had clinical presentations that were most consistent with MODY. In contrast, the ‘inconclusive’ and ‘inconsistent’ genes did not differ clinically from non‐carriers in diabetes‐related characteristics. CONCLUSIONS: Data from a large population‐based study provided novel statistical evidence to identify 6 MODY genes as consistent with being highly penetrant. These results have potential implications for interpreting genetic testing results and clinical diagnosis of MODY.