Fig. 1: An overview of MHCfovea.
From: Connecting MHC-I-binding motifs with HLA alleles via deep learning
MHCfovea, a deep learning-based framework, contains a prediction module and a summarization module that infers the relation between MHC-I sequences and peptide-binding motifs. First, the predictor, an ensemble model of multiple convolutional neural networks (CNN models), was trained on 150 observed alleles. In the predictor, 42 important positions were highlighted from MHC-I sequence (182 a.a.) using ScoreCAM. Next, we made predictions on 150 observed alleles and 12,858 unobserved alleles against a peptide dataset (number: 254,742) and extracted positive predictions (score >0.9) to generate the binding motif of an allele. Then, after clustering the N-terminal and C-terminal sub-motifs, we built hyper-motifs and the corresponding allele signatures based on 42 important positions to reveal the relation between binding motifs and MHC-I sequences.
