Predicting transcription factor binding sites with histone modification

Abstract

Histone modifications are associated with gene expression, influencing bind of transcription factor. Promoter and enhancer of human are characterized to combinatorial patterns of histone acetylation and methylation from ChIP-Seq data. However, because the data is only relative with gene, relation transcription factor of histone modification remains unclear. Here, we reported specific patterns of histone modification on transcriptional factor binding site (TFBS) from human CD4+ T cells. In result of comparing position of TFBS from TRANSFAC and position histone modification with gene expression, it represented a different histone pattern in TFBS between active and inactive gene expression. Specially, level of H3K4me3 at TFBS in active gene expression is a higher than silencing gene expression. This difference showed that we distinguished activity of TF binding in cell context. We also improved performance of TFBS prediction compared with sequence-based prediction. From this result, we developed new method for predicting TFBS to integrate sequence-based model with histone modification.