Mapping genotype-phenotype associations of nsSNPs in coiled-coil oligomerization domains of the human proteome.

Abstract

We assessed the impact of disease mutations versus polymorphisms in coiled-coil domains in Uniprot by modelling the structural and functional impact of variants in silico with the coiled-coil prediction program Multicoil. The structural impact of variants was evaluated with respect to three main metrics: the oligomerization score-to determine if the variant is stabilising or destabilising, the oligomerization state, and the register-specific score. The functional impact was queried indirectly in several ways. Firstly, we examined marginally stable coiled coils that were either stabilized or destabilized by the variant. Secondly, we looked for variants that altered the register of the wild-type coiled coil near wild-type irregularities of likely functional importance, such as skips and stammers. Thirdly, we searched for variants that altered the oligomerization state of the coiled coil. Disease mutations tended to be more destabilizing than polymorphisms; but interestingly, polymorphisms were more frequently associated with predicted changes in the oligomerization state. The functional impact was also queried by testing the association of coiled-coil variants with multiple phenotypes i.e. pleiotropy. Mutations in coiled-coil regions of proteins cause 155 different phenotypes and are more frequently associated with pleiotropy than proteins in general. Importantly, the coiled-coil region itself often encodes the pleiotropy. This article is protected by copyright. All rights reserved.