Figure 3.

Effects of sequence divergence on capture bias. Boxplots showed the distribution of capture bias (i.e. log2 transformed ratio between the number of reads derived from C57BL/6 J and that from SPRET/EiJ allele) for the targets without indels grouped by the number of SNVs (white boxes) based on exome sequencing (A), simulated exome sequencing data (B) or WGS (C). In exome sequencing, compared with the targets with identical sequence between the two genomes, the targets containing SNVs showed significantly higher capture biases toward C57BL/J. The more SNVs the targets contained, the higher the bias was observed. Compared with the targets containing no indels (white boxes), those containing the same number of SNVs, but also indels (grey boxes), showed significantly higher capture biases toward C57BL/J. Both biases were not observed in simulated and WGS data. * denotes statistically significant differences (P < 0.01).