Fig. 5

AT-rich genes have high diversity despite experiencing moderate mutation rates in the present. A, B Ratio of non-synonymous (missense plus loss of function) versus synonymous rare variants in the MANE gene set identified in gnomAD v2.1.1 exome sequencing of > 100,000 unrelated individuals [82]. Genes are binned by isochore decile (A) or k-means cluster (B), and dots indicate means. gnomAD rare variants are defined by < 0.1% allele frequency. C, D Raw counts of rare synonymous variants per gene in gnomAD v2.1.1 binned by isochore decile (C) or k-means cluster (D). E, F Variant rates from gnomAD v2.1.1 for genes with or without paralogous neighbors. Box plots show median and mid-quartile distribution. G, H Cumulative frequency distribution plots of gnomAD pLI (likelihood that a gene is loss-of-function intolerant) relative to a gene’s k-means cluster assignment (G) or promoter GC% (H) [82]. I, J Number of de novo point mutations observed per kb across the genes (TSS to TES) within an isochore relative to isochore GC% (I) and isochore GC% binned by decile (J). ~700,000 DNM calls are pooled from all ~11,000 trios sequenced to date [85]