Abstract
Understanding gene expression will require understanding where regulatory factors bind genomic DNA. The frequently used sequence-based motifs of protein-DNA binding are not predictive, since a genome contains many more binding sites than are actually bound and transcription factors of the same family share similar DNA-binding motifs. Traditionally, these motifs only depict sequence but neglect DNA shape. Since shape may contribute non-linearly and combinational to binding, machine learning approaches ought to be able to better predict transcription factor binding. Here we show that a random forest machine learning approach, which incorporates the 3D-shape of DNA, enhances binding prediction for all 216 tested Arabidopsis thaliana transcription factors and improves the resolution of differential binding by transcription factor family members which share the same binding motif. We observed that DNA shape features were individually weighted for each transcription factor, even if they shared the same binding sequence.
Highlights
Understanding gene expression will require understanding where regulatory factors bind genomic DNA
Changes in gene expression during development and invoked by environmental perturbations are critical to organismal function and these changes are influenced by DNA-binding transcription factors (TFs)
Arabidopsis thaliana encodes 1533 DNA-binding TFs1 many of which occur in protein families of a few to over a hundred members[2]
Summary
DNA shape features explain large part of protein–DNA binding affinity. To generate the datasets necessary for training, test, and validation, for each TF the sequence-based binding motif ( called “core motif”) was determined with MEMEChIP using all ampDAP-seq peaks. For ERF036, the slide at position -1 relative to the motif and the helix twist at position 5 in the motif is most influential, whereas for CBF4 the minor groove width at position 6 and the helix twist at position −1 contribute most to the decision of the RF model This observation underlines that the TFs, even though binding to the same core motif, are dependent on different peculiarities regarding the shape of the DNA (Fig. 2). For the HY5 (AT5G11260) TF of the bZIP family with the core motif ACGT, six DNA sequences with high (>150 peak height units) and low (
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
