Fig. 9: Factors that contribute to absence of support for deletions in HG002. | Nature Communications

Fig. 9: Factors that contribute to absence of support for deletions in HG002.

From: SurVIndel2: improving copy number variant calling from next-generation sequencing using hidden split reads

Fig. 9

a As expected, a large portion of unsupported deletions have a low expected support (ES) score (left). ES score is the probability that a read containing the breakpoint of the deletion will support its existence, i.e., it will be either a split or a hidden split read. On the other hand, most deletions with support have very high ES scores (right). However, a low ES score is not the only factor that contributes to the lack of support: more than 40% of the unsupported deletions have an ES score ≥ 0.5. b Another factor is a lack of correctly sequenced reads containing the breakpoint of the deletions. For each deletion with a high-quality alternative allele (“Methods”), we mapped the short reads of HG002 to the alternative allele and counted the reads correctly sequenced (i.e., without mismatches or indels) and containing the deletion breakpoint. We further stratified the deletions into low ES  (< 0.3), medium ES (≥ 0.3 and  < 0.5) and high ES (≥ 0.5). High ES unsupported deletions (left) tend to have very few, or no reads correctly sequenced that contain the breakpoint, while deletions with a higher number of reads containing the breakpoint have low ES. Supported deletions (right), on the other hand, have both ingredients necessary for support: reads containing the breakpoint, and high ES.

Back to article page