Abstract
Accurate computational identification of promoters remains a challenge as these key DNA regulatory regions have variable structures composed of functional motifs that provide gene-specific initiation of transcription. In this paper we utilize Convolutional Neural Networks (CNN) to analyze sequence characteristics of prokaryotic and eukaryotic promoters and build their predictive models. We trained a similar CNN architecture on promoters of five distant organisms: human, mouse, plant (Arabidopsis), and two bacteria (Escherichia coli and Bacillus subtilis). We found that CNN trained on sigma70 subclass of Escherichia coli promoter gives an excellent classification of promoters and non-promoter sequences (Sn = 0.90, Sp = 0.96, CC = 0.84). The Bacillus subtilis promoters identification CNN model achieves Sn = 0.91, Sp = 0.95, and CC = 0.86. For human, mouse and Arabidopsis promoters we employed CNNs for identification of two well-known promoter classes (TATA and non-TATA promoters). CNN models nicely recognize these complex functional regions. For human promoters Sn/Sp/CC accuracy of prediction reached 0.95/0.98/0,90 on TATA and 0.90/0.98/0.89 for non-TATA promoter sequences, respectively. For Arabidopsis we observed Sn/Sp/CC 0.95/0.97/0.91 (TATA) and 0.94/0.94/0.86 (non-TATA) promoters. Thus, the developed CNN models, implemented in CNNProm program, demonstrated the ability of deep learning approach to grasp complex promoter sequence characteristics and achieve significantly higher accuracy compared to the previously developed promoter prediction programs. We also propose random substitution procedure to discover positionally conserved promoter functional elements. As the suggested approach does not require knowledge of any specific promoter features, it can be easily extended to identify promoters and other complex functional regions in sequences of many other and especially newly sequenced genomes. The CNNProm program is available to run at web server http://www.softberry.com.
Highlights
Promoter is a key region that is involved in differential transcription regulation of proteincoding and RNA genes
We face the situation that specific promoter characteristics that are often used in developing promoter predictors are poorly understood in many new genomes. This creates favorable circumstances for developing universally applicable algorithm of promoter prediction and in this paper we propose the use of convolutional neural networks, with an input consisting of only genomic sequence, as a rather general approach to solution of this problem
The developed Convolutional Neural Networks (CNN) models, implemented in CNNProm program, demonstrated the ability of deep learning to grasp complex promoter sequence characteristics and achieve significantly higher accuracy compared to previously developed promoter prediction programs
Summary
Promoter is a key region that is involved in differential transcription regulation of proteincoding and RNA genes. Promoter 5’-flanking regions may contain many short (5–10 bases long) motifs that serve as recognition sites for proteins providing initiation of transcription as well as specific regulation of gene expression. About 30–50% of all known eukaryotic promoters contain a TATA-box at a position *30 bp upstream from the transcription start site. Large groups of genes including housekeeping genes, some oncogenes and growth factor genes possess TATA-less promoters. In these promoters Inr (the initiator region) or the recently found downstream promoter element (DPE), usually located *25–30 bp downstream of TSS, may control the exact position of the transcription start [1, 2]
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
