Frequency of ACMG-56 Variants in Whole Genomes of Healthy Elderly. L. Ariniello1, C. S. Bloss1, G. Erickson1, P. Pham2, D. Boeldt1, O. Libiger1, N. Schork1, E. Topol1, A. Van Zeeland2, A. Torkamani1 1) Scripps Translational Science Institute, La Jolla, CA; 2) Cypher Genomics, San Diego, CA.

   Last year the American College of Medical Genetics (ACMG) recommended that laboratories performing clinical sequencing seek and report mutations of known or expected pathogenic mutations across a list of 56 genes implicated predominately in cancer and cardiovascular disease risk. It was recommended that this be performed for all clinical germline exome and genome sequencing, irrespective of patient age. Unfortunately, there are few studies on the collective frequency of mutations across this list of 56 genes in any given population. Such information could enable laboratories to anticipate their potential reporting burden. In the current study, we aimed to determine the frequency of mutations across the ACMG-56 in the Scripps Wellderly Cohort. The Wellderly cohort is comprised of individuals age 80 or older without a history of any major chronic diseases or medication use. Whole genome sequencing data generated by Complete Genomics for a subset of N=454 (mean age=86.9, range 80-104 years) unrelated European individuals were analyzed. Variants were extracted per ACMG-56 guidelines using the Cypher Genomics annotation pipeline, which returns both reported pathogenic variants as well as predicted pathogenic variants based on allele frequency in various reference populations and algorithmic predictions of functional impact. Across the sample, 609 non-unique variants were identified. A total of 314 individuals (69.2%) were heterozygous for at least one variant, and a range of 0 to 6 variants per individual were identified (mean=1.34, SD=1.27). Six Wellderly individuals (1.3%) were homozygous for one variant; however, further analysis of these instances of homozygosity revealed questionable evidence of pathogenicity for the variants identified. There are few published studies of the frequency of ACMG-56 variants across a population. Dulik et al. (2013) reported a prior analysis of a different cohort, unselected for a healthy aging phenotype, in which a range of 1 to 6 (mean=3) variants per individual were identified. Frequencies in the Scripps Wellderly cohort were lower, which would be expected given the selection criteria for this cohort. The ACMG-56 recommendations have been highly controversial for a number of reasons, and although many factors may influence laboratories adoption of the guidelines, anticipating the frequency of variants likely to be identified, and thus the reporting burden, may help inform such decisions.

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