Individualized iterative phenotyping for genome-wide analysis of loss of function mutations. J. J. Johnston1, K. Lewis1, D. Ng1, L. N. Singh1, J. Wynter1, C. Brewer2, B. P. Brooks3, I. Brownell4, F. Candotti1, S. G. Gonsalves1, P. S. Hart1, H. H. Kong4, K. I. Rother5, R. Sokolic1, B. D. Solomon1, W. M. Zein3, D. N. Cooper6, P. D. Stenson6, J. C. Mullikin1, 7, L. G. Biesecker1, 7 1) National Human Genome Research Institute, National Institutes of Health, Bethesda, MD, USA; 2) National Institute on Deafness and other Communicative Disorders, National Institutes of Health, Bethesda, MD, USA; 3) National Eye Institute, National Institutes of Health, Bethesda, MD, USA; 4) National Cancer Institute, National Institutes of Health, Bethesda, MD, USA; 5) National Institute of Diabetes and Digestive Diseases, National Institutes of Health, Bethesda, MD, USA; 6) Institute of Medical Genetics, School of Medicine, Cardiff University, Heath Park, Cardiff, United Kingdom; 7) NIH Intramural Sequencing Center, National Institutes of Health, Bethesda, MD, USA.

   Putative loss of function (pLOF) variants are common in genomes and it is critical for understanding gene function and predictive medicine to assess the consequences of pLOF variants. Towards this end, we characterized the consequences of pLOF variants in a largely healthy adult exome cohort by iterative phenotyping. Exome data were generated on 951 participants from the ClinSeq cohort and filtered for pLOF variants likely to cause a phenotype in heterozygotes. 106 of 951 exomes had such a pLOF variant and 79 participants were available for evaluation. Of those 79, 38 had positive findings or family histories (31 variants in 19 genes), two had indeterminate findings (two variants, one each in two genes) and 39 had negative findings or a negative family history (32 variants in 27 genes) for that trait. We correlated these findings with the haploinsufficiency score of the affected gene when available, the Combined Annotation-Dependent Depletion (CADD) score of the variant, and a novel variant composite pathogenicity score. The composite pathogenicity score was calculated based on the following seven attributes of each variant: whether the variant predicted a frameshift with >10 aberrant amino acids, whether MutationTaster predicted nonsense-mediated decay, whether 5 pLOF mutations were present for the gene in the Human Gene Mutation Database (HGMD), whether there was a mutation in that exon in the public version of HGMD, whether that specific mutation was in HGMD as a disease causing (DM) mutation, whether the pLOF was in the middle 90% of the gene (i.e., not in the first or last 5%), and whether the exon with the pLOF variant was an exon in the dominant gene model as defined by presence in >75% of overlapping spliced ESTs. Using these measures we developed an algorithm to classify variants. We conclude that 1/30 unselected individuals had a manifest phenotype attributable to rare LOF variants, which is more common than may be assumed.

You may contact the first author (during and after the meeting) at