Genetic association analysis of 269 rare diseases reveals novel aetiologies

Abstract

The genetic aetiologies of more than half of rare diseases remain unknown1. Standardised genome sequencing (GS) and phenotyping of large patient cohorts provides an opportunity for discovering the unknown aetiologies2, but this depends on efficient and powerful analytical methods3. We have developed a portable computational and statistical framework for inferring genetic associations with rare diseases. At its core lies the Rareservoir, a compact database of rare variant genotypes and phenotypes. We built a Rareservoir of 77,539 genomes sequenced by the 100,000 Genomes Project (100KGP)4. We then applied the Bayesian association method, BeviMed3, across 269 rare diseases assigned to participants in the project, identifying 238 known5 and 21 novel associations. The novel results included three which we selected for validation. We provide compelling evidence that (1) loss-of-function variants in the ETS-family transcription factor encoding gene ERG lead to primary lymphoedema, (2) truncating variants in the last exon of TGF{beta} regulator PMEPA1 result in Loeys-Dietz syndrome6, and (3) loss-of-function variants in GPR156 give rise to recessive congenital hearing impairment. These novel findings confirm the power of our analytical approach for the aetiological discovery of rare diseases.