We compared the rate of de novo mutation per base pair (bp), restricting these analyses to base pairs covered at ≥20× in all members of a trio (our minimum criteria for de novo calling; Sanders et al., 2012). We calculated the rate per base pair as the total number of variants observed within this target region. More specifically, for the overall (coding plus non-coding) mutation rate (e.g., Table 2), this encompassed the exome capture array intervals plus the 100 bp of interval padding added during GATK processing (denoted as “total callable” in Table 1). For coding mutation rate (e.g., Table 2; Figures 2A, 2B, 3A, and 3B), this encompassed the intersection of these intervals with the coding portion of the exome based on RefSeq hg19 gene definitions (“total callable exome” in Table 1). This strategy normalizes for differences in capture array design and coverage distribution across the exome.
