Fig. 3

The BIND&MODIFY disclose the H3K27me3 distribution around pericentromeres. A We simulated the reads with different length based on the pericentromeric regions (PE50-paired end sequencing 50 bp, SE50-single end sequencing 50 bp) (Methods). The pericentromeric region is defined as the boundary regions of the centromeres which could be labeled by H3K27me3 (The center areas of centromeres cannot be labeled; reference genome: CHM13). The multiple alignment rates decreased with the increasing read length. The PE150 still have 25% multiple alignments in the pericentromeric regions. B In the real sequencing data, the SE50 demonstrated higher multiple alignment rates than the simulated SE50, due to the randomness size distribution. The BIND&MODIFY have the much lower multiple alignment rate, which could offer the more accurate epigenomic locations. C The H3K27me3 signal distribution on the pericentromeric regions (Chr20:26,260,000 ~ 26,320,000). The H3K27me3 signal increased to the highest on the boundary region of the centromere