Transcription factor binding site detection using convolutional neural networks with a functional group-based data representation

Abstract Transcription factors (TFs) play an essential role in molecular biology by regulating gene expression. The binding sites of TFs can vary a large amount and the numerous possible locations make their detection challenging issue. Recently, several machine learning approaches using nucleotide sequence data were applied to classify DNA sequences regarding Factor Binding Sites (TFBS). We propose novel training strategy without traditional 1D nucleotide-based representation instead 2D topological matrix sub-nucleotide chemical functional groups substantially defining protein ability fragments. train convolutional neural networks this Functional Group Representation (FGDR) solve TFBS classification task. compare our results with efficiency previous show that from FGDR has benefits classification. Moreover, we reason deep produces competitive while only introducing pre-processing conversion step. Finally, employing ensemble models representations for network higher performance than any single input approaches.