The population frequency of predicted pathogenic genetic variants in commonly affected CAKUT genes in the general population

BACKGROUND: Renal imaging suggests that congenital anomalies of the kidney and urinary tract (CAKUT) affect one in 200 of the population, and 20% have a monogenic cause. This study determined the population frequency of predicted pathogenic variants in six commonly affected CAKUT genes. METHODS: HNF1B, SALL1, EYA1, PBX1, GATA3, and PAX2 variants were downloaded from gnomAD v.2.1.1 (n = 141,456) and the population frequency of predicted disease-causing variants calculated from the sum of structural, null, and predicted pathogenic missense changes in the overall cohort or from variants shared with ClinVar, HGMD, or LOVD. Population frequencies were also determined in constituent ancestries and, using our method and ClinVar assessments, in a replication cohort (gnomAD v.4.1, n = 807,162). RESULTS: The population frequency of disease-causing variants in these six genes in CAKUT lies between one in 249 (our strategy) and one in 1263 (ClinVar assessments) in the gnomAD v.2.1.1 database. More than half these variants were missense changes. Predicted pathogenic variants were commonest in African-Americans (one in 149) and least common in Ashkenazim (one in 864). The population frequency estimated from gnomAD v.4.1 lies between one in 372 (our strategy) and one in 1,762 (with ClinVar). CONCLUSIONS: These calculations suggest that monogenic causes of CAKUT due to variants in these six genes are likely more common than the previously calculated one in 1,000. The ClinVar results are underestimates since assessments were not available for structural, copy number and many missense changes. However, some disease-causing variants identified here will not result in clinical disease because of incomplete penetrance and variable expressivity.