In order to obtain an overview of the distribution of regulatory variants, we annotated the SNVs of GM12878 from Complete Genomics with RegulomeDB (Fig. 2C,D) and mapped them onto genes and noncoding regions using the GENCODE v7 gene information (available from UCSC Table Browser) (Harrow et al. 2012). We find that a majority of variants are noncoding, with only 77,214 (2.1%) falling within exons. The coding variants can be further classified as those likely to cause damaging mutations in genes as predicted using PolyPhen-2 (Fig. 3; Supplemental Fig. S1; Adzhubei et al. 2010). In total, 2457 coding variants (0.07%) are predicted to cause damaging mutations in genes (PolyPhen-2 prediction of probably or possibly damaging). Of these, 626 (0.01%) are homozygous and likely to affect the gene product of both alleles.
