Genomic analyses in 101,127 UK women show that previously reported monogenic genes are not common causes of premature ovarian insufficiency

Abstract Study question What is the penetrance of variants in previously reported genes for premature ovarian insufficiency (POI) in the general population? Summary answer Heterozygous genetic variants in previously reported monogenic genes are not a common, highly penetrant cause of POI. What is known already Premature ovarian insufficiency (POI), defined as menopause before 40, has been associated with variants in over 70 genes. However, the evidence to support the causality of individual genetic variants varies between studies. As genomic data becomes more accessible, it is essential to ascertain the penetrance of variants in the absence of family history information. We identified 75 genes associated with POI from the literature, including family segregation studies, consanguineous pedigrees and cohort analyses of whole-exome/targeted sequencing data. Functional evaluation was also available for many of the genes. Study design, size, duration We used data from 101,127 females of European ancestry in the UK Biobank, to study the role of previously reported monogenic causal genes on ovarian function. We tested the association of 301 previously reported variants with POI and ANM, plus more than 2.5 million rare variants which were annotated and had not been implicated in POI previously. Genomic variants were tested individually and also combined into a gene burden test. Participants/materials, setting, methods Age at natural menopause (ANM) was derived from self-reported questionnaire data from the age at last menstrual period, excluding those with surgical menopause or taking hormone replacement therapy. POI cases were classified as women with ANM under 40 years (N = 2,213). Linked primary care records were used to identify a clinical cohort of POI cases, including primary amenorrhea (N = 113). Main results and the role of chance All of the previously reported individual variants we identified in our POI cases were also detected in the control group. Moreover, the gene burden tests were not associated with POI or ANM as a quantitative trait. Our results indicate that autosomal dominant causes of POI are rare; a single predicted loss-of-function (LOF) or non-synonymous genetic variant in one of the previously reported genes is generally not pathogenic. Limitations, reasons for caution The penetrance of monogenic disease-causing variants is likely to be lower in population-based cohorts such as UK Biobank than in clinically-ascertained cohorts. Wider implications of the findings Heterozygous LOF or non-synonymous variants in previously reported POI genes should be interpreted with caution and are unlikely to cause POI. The findings have implications for clinicians diagnosing causes of POI. Trial registration number Not applicable