Multi-Branch-CNN: Classification of ion channel interacting peptides using multi-branch convolutional neural network

Jielu Yan, Bob Zhang, Mingliang Zhou, Hang Fai Kwok, Shirley W. I. Siu

Abstract: Highlights•A deep learning method was proposed to improve ion channel peptide prediction.•CNN with multiple branches combined learning of different sequence representations.•Multi-Branch-CNN is superior to conventional machine learning and ensemble methods.•Eliminating overrepresented sequences can generate more reliable models.