DOI: 10.1101/468363Nov 14, 2018Paper

ACME: Pan-specific peptide-MHC class I binding prediction through attention-based deep neural networks

BioRxiv : the Preprint Server for Biology
Yan HuJianyang Zeng

Abstract

Prediction of peptide binding to MHC molecules plays a vital role in the development of therapeutic vaccines for the treatment of cancer. Although numerous computational methods have been developed to this end, several challenges still remain in predicting peptide-MHC interactions. Many previous methods are allele-specific, training separate models for individual alleles and are thus unable to yield accurate predictions for those alleles with limited training data. Despite that there exist several pan-specific algorithms that train a common model for different alleles, they only adopt simple model structures that generally have limited performance in capturing the complex underlying patterns of peptide-MHC interactions. Here we present ACME (Attention-based Convolutional neural networks for MHC Epitope binding prediction), a new pan-specific algorithm to accurately predict the binding affinities between peptides and MHC class I molecules, even for those new alleles that are not seen in the training data. Extensive tests have demonstrated that ACME can significantly outperform other state-of-the-art prediction methods with an increase of the Pearson Correlation Coefficient by up to 23 percent. In addition, its ability to identif...Continue Reading

Related Concepts

Alleles
Epitopes
Attention
Malignant Neoplasms
Histocompatibility Antigens Class I
Peptides
Oxytocin, N-acetyl-2-O-methyl-Tyr-
Therapeutic vaccine, NOS
Binding (Molecular Function)
Small Molecule

Related Feeds

BioRxiv & MedRxiv Preprints

BioRxiv and MedRxiv are the preprint servers for biology and health sciences respectively, operated by Cold Spring Harbor Laboratory. Here are the latest preprint articles (which are not peer-reviewed) from BioRxiv and MedRxiv.