Mapping a genetic cause of hypermutation in 26537

To identify a genetic cause for the hypermutation phenotype of 26537, we used two approaches. First, we used mGAP, the macaque genotype and phenotype database, to annotate allele frequencies for all variants in the joint VCF of individuals in the study. We then selected all coding variants from 26537 with mGAP minor allele frequency < 1%, homozygous in 26537, and not homozygous in the other trio parents. Sequencing alignments and gene annotations were then used to curate the resulting list, eliminating sequencing errors and annotation errors. Of the remaining variants, the only variant that affected a gene with a known function relating to DNA repair was MBD4.

Second, to confirm MBD4 as the most likely cause of the hypermutation phenotype, we carefully investigated 1378 candidate hypermutation genes (Supplemental Tables S3–S6), chosen based on previous literature tying them to genome instability. These are cancer predisposition genes in pediatric cancer (Zhang et al. 2015), DNA damage repair genes from The Cancer Genome Atlas (Knijnenburg et al. 2018), an in-house cancer driver list (“Tarpey driver genes”), cancer gene consensus genes from COSMIC (https://cancer.sanger.ac.uk/census), and genes identified as involved in developmental disorders (“DDD_genes”; https://www.deciphergenomics.org/ddd/ddgenes). The full list of genes interrogated is provided in Supplemental Table S3. The joint VCF of germline genotypes was annotated using VEP. We then extracted all mutations in these genes that were annotated by VEP as “high” or “moderate” impact, and identified gene candidates with biallelic damaging mutations in at least one dam or sire, which generated a short list of 23 genes (Fig. 4).