Discriminating coding from non-coding regions based on codon structure and methylation-mediated substitution: An application in rice and cattle

Coding regions are the fragments of DNA sequence that codes for protein through the process of transcription and translation respectively. On the other hand, the non coding regions do not give rise to any protein. Discrimination of coding regions from the non coding regions is essential for genome annotation. In this study, an attempt has been made to develop a random forest based computational approach for discriminating coding regions (CDS) from non-coding regions (introns). The features based on codon structure and methylation mediated substitutions were used in this approach. The developed approach achieved high classification accuracy, while tested on two agriculturally important species i.e., rice and cattle. The proposed approach is believed to complement the other prediction methods. Based on the proposed approach, an online prediction server 'DCDNC' has also been developed for easy prediction by the users. The prediction server is freely available at http://cabgrid.res.in:8080/DCDNC.