Link to bioRxiv paper:
http://biorxiv.org/cgi/content/short/2023.02.16.528799v1?rss=1
Authors: Koyama, K., Hashimoto, K., Nagao, C., Mizuguchi, K.
Abstract:
Motivation: Understanding how a T cell receptor (TCR) recognizes its specific ligand peptide is crucial for gaining insight into biological functions and disease mechanisms. Despite its importance, experimentally determining TCR-peptide interactions can be expensive and time-consuming. To address this challenge, several computational methods have been proposed, but none have incorporated an attention layer from language models in conjunction with TCR-peptide "structural" binding mechanisms. Results: In this study, we developed a machine learning model based on a modified version of the transformer, a cross-attention neural network, to predict TCR-peptide binding solely based on protein sequences. This model achieved state-of-the-art performance as measured by the ROCAUC score on a benchmark dataset of TCR-peptide binding. Furthermore, by analyzing the results of binding predictions, we associated the neural network weights with protein structural properties. By classifying the residues into large and small attention groups, we identified statistically significant properties associated with the largely attended residues, such as hydrogen-bond connections within the TCR. This ability to provide an interpretable prediction of TCR-peptide binding should increase our knowledge of molecular recognition and pave the way to designing new therapeutics.
Copy rights belong to original authors. Visit the link for more info
Podcast created by Paper Player, LLC
