Estimating human mutation rate using autozygosity in a founder population. C. D. Campbell1, J. X. Chong2, M. Malig1, A. Ko1, B. L. Dumont1, L. Han2, L. Vives1, B. J. O'Roak1, P. H. Sudmant1, M. Abney2, C. Ober2, E. E. Eichler1,3 1) Department of Genome Sciences, University of Washington, Seattle, WA; 2) Department of Human Genetics, The University of Chicago; 3) Howard Hughes Medical Institute.

   Understanding the rate and pattern of new mutation is critical to our understanding of human disease and evolution. Direct estimates have been complicated by the relatively modest number of events per generation, difficulties in distinguishing somatic (including cell line-derived) from germ-line mutations, and biases in targeted capture and resequencing. We took advantage of the extensive autozygosity (i.e. homozygosity by recent decent) in a genealogically well-defined population of Hutterites to estimate the mutation rate over multiple generations, since heterozygous mutations in these regions are necessarily recent in origin. We sequenced whole genomes from whole blood DNA of 15 individuals from five parent-offspring trios. Based on comparisons to genomes from other populations, we observed a 33-fold increase in autozygous basepairs in segments greater than 2 Mbp in Hutterite individuals. We focused on 44 segments of autozygosity greater than 5 Mbp in the Hutterites, since segments of this size were absent in the genomes of non-Hutterite individuals. Using genotyping data from the extended Hutterite pedigree, we computed the number of meioses separating each pair of autozygous alleles and determined the number of heterozygous single nucleotide variants (SNVs) in each segment. We validated 72 SNVs from 498 Mbp of autozygous DNA providing an SNV mutation rate of 1.21x10-8 (95% confidence interval 1-1.5x10-8) mutations per basepair per generation. We observed a nine-fold increase in the mutation rate for bases within CpG dinucleotides (10.3x10-8 mutations per CpG basepair and per generation) and strong evidence (p = 0.002) for a paternal bias in the origin of new mutations. In addition, we found evidence of a long-range gene conversion event (at least 6 kbp) occurring within six meioses at the C4, TNX locus on chromosome 6. Interestingly, even after correcting for CpG bias, we observed a nonrandom distribution of heterozygous SNVs (both novel and known) in the autozygous segments (p = 0.001) suggesting mutational hotspots or sites of long-range gene conversion.