Fig. 1

Strand-specific amplification of non-targeted sequences at poly(T/A) sites in the SMART ChIP-seq analysis. a Flowchart of the SMART ChIP-seq procedure at non-poly(T/A) sites, adapted from the user manual of the kit (https://www.takarabio.com/documents/User%20Manual/DNA%20SMART%20ChIP-Seq%20Kit%20User%20Manual_101617.pdf). b, c Modified flowcharts to show annealing of the SMART poly(dA) primers to non-tailed Ts within targeted (b) or non-targeted (c) DNA templates, leading to strand-specific amplification at poly(T) sites. For poly(A) sites, false amplification occurs to the opposite strand. d-f ChIP-seq read densities at three randomly picked non-poly(T/A) and poly(T/A) sites. The data is from SRR3229031 (Additional file 1: Table S1, Dataset 1), and Integrative Genomics Viewer (IGV) [32] is used to show the ChIP-seq reads from paired-end (PE) or single-end (SE) sequencing. For PE, read1 and read2 are shown as pairs, with reads mapped to “+” and “-“strands in red and blue, respectively. For SE, only Read1 (extracted from PE data) is shown. g-i. Aggregated read distribution at non-poly(T/A) and poly(T/A) sites. In h and i, poly(T/A) sites were defined as those with ≥12 consecutive T or A in the human reference genome. To define non-poly(T/A) sites, we first selected genomic regions that are > 4 kb in length and > 1 kb away from poly(T/A) sites, and then take the 2 kb regions around the middle points. In total, we got 301,474 non-poly(T/A) sites, 338,568 poly(T) sites, and 336,703 poly(A) sites. Refer to the Method section (SE mode, Additional file 2: Fig. S6) for the calculation of read distribution