Fig. 2: CGN arginine codon frequency and gene pathogenicity.

(a) CGN arginine codon frequency comparison between syndromic SFARI genes and control genes not linked to autism spectrum disease (ASD) in SFARI. (b) CGN arginine codon frequency comparison between human homologs to essential genes detected in Drosophila X-screen and all other genes (controls). (c) CGN arginine codon frequency comparison between OMIM disease genes and nondisease gene controls. (d–f) CGN codon frequency comparison with gnomAD constraint metrics: probability of loss-of-function intolerance (pLI) score, observed-over-expected (o/e) ratio of loss-of-function (LoF) and o/e ratio of missense variants. (g) CGN codon frequency comparison with GC content 5, 10, 25, and 50 kb up- and downstream of transcript boundaries. Pearson correlation coefficients were r_{5 kb} = 0.537, r_{10 kb} = 0.532, r_{25 kb} = 0.519, and r_{50 kb} = 0.503. (h–i) Examples of median gene transcripts per million (TPM) for cerebellum and left ventricle.