Zhonghao Liu

DeepSeqPanII predicts peptide-HLA class II binding using an interpretable recurrent neural network with attention mechanisms, achieving state-of-the-art performance.

The paper introduces DeepAttentionPan, a deep learning model for MHC-I peptide binding prediction that uses attention mechanisms to improve interpretability and performance.

DeepSeqPan is a deep convolutional neural network for predicting HLA-peptide binding affinity without requiring structural information, achieving state-of-the-art performance on multiple HLA alleles.

DeepMHC uses deep convolutional neural networks to predict MHC-I peptide binding affinity, outperforming conventional methods without requiring additional features.