Proposed Cover Art
Deep neural networks identify sequence context features predictive of transcription factor binding
The transcription process of DNA is highly complex and while short DNA sequence motifs recognized by transcription factors are well known, less is known about the context in the DNA sequence that determines whether a transcription factor will actually bind its motif. This paper presents a method that uses convolutional neural networks to identify sequence features that help predict whether transcribing proteins can bind to their target sequences in DNA
The proposed illustration depicts a stylized DNA strand where some proteins have bound and others have not, according to the prediction model, and the "activity" of one motif along the DNA.