H3K4me2 reliably defines transcription factor binding regions in different cells

- On average about 90% of transcription factor binding regions (TFBRs) overlap with the H3K4me2 regions.
- H3K4me2 regions with stronger H3K4me2 signals more likely overlap with TFBRs than those with weaker signals.
- H3K4me2 regions together with H3K27ac regions can greatly reduce false positive predictions of TFBRs.

Histone modification (HM) patterns are widely applied to identify transcription factor binding regions (TFBRs). However, how frequently the TFBRs overlap with genomic regions enriched with certain types of HMs and which HM marker is more effective to pinpoint the TFBRs have not been systematically investigated. To address these problems, we studied 149 transcription factor (TF) ChIP-seq datasets and 33 HM ChIP-seq datasets in three cell lines. We found that on average about 90% of the TFBRs overlap with the H3K4me2-enriched regions. Moreover, the H3K4me2-enriched regions with stronger signals of H3K4me2 enrichment more likely overlap with the TFBRs than those with weaker signals. In addition, we showed that the H3K4me2-enriched regions together with the H3K27ac-enriched regions can greatly reduce false positive predictions of the TFBRs. Our study sheds light on the comprehensive discovery of the TFBRs using the HeK4me-enriched regions, especially when no good antibody to a TF exists.