DNA breathing and shape encode distinct mechanisms of transcription factor binding

Abstract Motivation Interactions between transcription factors (TFs) and DNA play a key role in regulating gene expression. While these interactions are often viewed primarily as recognition of DNA sequence motifs, several studies have shown the limitation of this approach and highlighted the impact of physical properties of DNA. Prior studies have supported the importance of DNA shape features as determinant of TF–DNA recognition. Another physical property is DNA breathing, the spontaneous and transient opening of double-stranded DNA through thermal motions, which has not been systematically examined. Results We analyzed in vitro TF–DNA binding data and found that DNA breathing features in motif regions are correlated with binding affinity, highlighting a role of locally and temporally melted DNA. Here, we compared contributions of DNA breathing and shape to TF binding. These analyses suggest that DNA breathing provides sequence-dependent information related to DNA structure not fully captured by DNA shape. Extending the analysis to ChIP-seq data, we found that breathing features are associated with TF binding in vivo, although the direction and magnitude of these associations vary across TFs. Our results support DNA breathing as an additional physical feature contributing to TF–DNA interactions alongside DNA sequence motifs and DNA shape. Availability and implementation Data and analysis code are available at https://doi.org/10.17605/OSF.IO/UAHBJ and https://github.com/TsuPeiChiu/DNAbreathing.