Improved Predictions of MHC-Peptide Binding using Protein Language Models

Abstract

Major histocompatibility complex (MHC) molecules bind to peptides from exogenous antigens, and present them on the surface of cells, allowing the immune system (T cells) to detect them. Elucidating the process of this presentation is essential for regulation and potential manipulation of the cellular immune system. Predicting whether a given peptide will bind to the MHC is an important step in the above process, motivating the introduction of many computational approaches. NetMHCPan, a pan-specific model predicting binding of peptides to any MHC molecule, is one of the most widely used methods which focuses on solving this binary classification problem using a shallow neural network. The successful results of AI methods, especially Natural Language Processing (NLP based) pretrained models in various applications including protein structure determination, motivated us to explore their use in this problem as well. Specifically, we considered fine-tuning these large deep learning models using as dataset the peptide-MHC sequences. Using standard metrics in this area, and the same training and test sets, we show that our model outperforms NetMHCpan4.1 which has been shown to outperform all other earlier methods.

Machine Learning Approaches to TCR Repertoire Analysis

“…Therefore, the models are not evaluated in terms of which downstream tasks can be applied via transfer learning. Recently, attempts appear, which utilize large language models in repertoire analysis (133)(134)(135)(136)(137)180). In AntiBERTa (137), fine-tuning for a downstream task is also investigated.…”

“…ImmunoBERT (134) used the same pre-trained model for the peptide-MHC (Class I) binding prediction task. Hashemi et al (135) employed the pre-trained model of ( 131) and fine-tuned them for peptide-MHC (Class I) binding prediction and achieved higher performance compared to a previous software. Some papers perform pre-training on their own on the repertoire sequencing dataset.…”

A Computational Pipeline for Accurate Prioritization of Protein‐Protein Binding Candidates in High‐Throughput Protein Libraries

“…As the field advances, we expect to see better AI methods that coupled with this strategy, will lead to better predictions. [53,54] Finally, as interest in peptides as potential drugs increases, so does the ability of AI software to handle peptide cyclizations, [55,56] which results in improved membrane permeability and increased resistance to degradation.…”

Assessing developability early in the discovery process for novel biologics

“…Use of multi-scale molecular simulations and machine learning models at these stages can help with identifying early formulation process challenges, such as aggregation, diffusion interaction, viscosity and solubility, 94–104 physicochemical degradation, 105–107 and immunogenicity. 108 , 109 Specifically, use of explicit solvent molecular dynamics (MD) simulations can potentially provide a molecular-level understanding of molecular response to thermal and other stresses. 110 Expanding the scope of such simulations to include the considerations of formulation buffers, salt, pH, and excipients will pave the way toward in silico formulation development for biologics.…”