Long-read sequencing improves sensitivity to discover variation in complex repetitive regions, assign parent-of-origin, and distinguish germline from postzygotic mutations. We applied Illumina, Oxford Nanopore Technologies, and PacBio sequencing to discover and validate de novo mutations in 73 children from 42 autism families (157 individuals). We assay 2.77 Gbp of the human genome, yielding on average 95 de novo mutations per transmission (87.5 single-nucleotide substitutions, 7.8 indels), with no significant difference in mutation rate or profile between probands and their unaffected siblings. Long reads increase de novo mutation discovery by 20-40% and double the mutations classified as early embryonic. The germline mutation rate is 1.30×10-8 substitutions/base pair/generation; the postzygotic rate is 0.23×10-8. These rates are significantly increased in repetitive DNA, where segmental duplication mutability is dependent on length and percent identity. Here, we show that enrichment in repeats occurs predominantly postzygotically, likely resulting from faulty DNA repair and interlocus gene conversion.
Noyes et al. (Mon,) studied this question.