EPI-PRED: Leveraging CNN and Bi-LSTM Models for Accurate Prediction of Enhancer-Promoter Interactions

Abstract

The identification of Enhancers-promoter interactions (EPI’s) will assist in understanding the genetic regulation mechanisms. EPIs are determined through time-consuming and laborious testing techniques.  Several methods are being contributed to deal with this issue. Due to their promising ability to predict, DL-based techniques have been extensively employed in the genome-scale detection of EPIs recently. This study is to employ the  CNN and BiLSTM model named as “EPI-PRED”  in predicting the EPI’s with a set of features trained using the DL models in a python platform and it is successful predicted and  evaluated with the SEPT, EPI-Trans and TF-EPI models using sensitivity, Specificity , precision and accuracy . It outperforms the other state of art DL methods and offers a new avenues in medical research.