Extended Data Fig. 2: Defining the CN-mappable genome.
From: Most large structural variants in cancer genomes can be detected without long reads
(a) Example of read depth in the matched normal sample of TCGA case 6864 at a locus with a high density of unmappable bases (left, CN-unmappable) and at a locus with low density of unmappable bases (right, CN-mappable). In both panels, the top track shows purity / ploidy transformed read depth per 1 kbp bin (scaled to CN units) and the bottom track shows the density of unmappable bases per 10 kbp bin. The expected purty / ploidy transformed read depth for all bins is two as this is a non-neoplastic sample. (b) Fraction of the genome classified as CN-unmappable (top), base-unmappable (middle), and fully-mappable (bottom). CN-unmappable regions refer to genomic positions surrounded by ≥90% multi-mapping bases (bases with multiple alignments at a length of 101 bp and > 96% homology) in their 1 kbp vicinity. Base-unmappable regions refer to bases pairs falling within CN-unmappable regions, in addition to base pairs in lower density regions that are multi-mapping. The remainder of the genome is comprised of fully-mappable bases. SINE, short interspersed nuclear element. LINE, long interspersed nuclear element. LCR, low copy repeat, (c) Distribution of the standard deviation of read depth within 10 kbp bins for normal samples in CN-unmappable, base-unmappable, and fully-mappable regions in normal samples (n = 1000 CN-unmappable, 1000 base-unmappable, 1000 fully-mappable tiles sampled across 100 normal short-read whole genome sequencing profiles). Box plot: line (median), body (IQR), whiskers (1.5 times IQR). (d) The fraction of the genome designated CN-unmappable, base-unmappable, and fully-mappable using either 101 bp or 150 bp self-alignment, and either GrCh37 or GrCh38 as the reference genome. The fraction of the genome that is CN-unmappable is 13% in all three cases shown.
