English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Motivation: The activity of the adaptive immune system is governed by T-cellsand their specific T-cell receptors (TCR), which selectively recognize foreignantigens. Recent advances in experimental techniques have enabled sequencing ofTCRs and their antigenic targets (epitopes), allowing to research the missinglink between TCR sequence and epitope binding specificity. Scarcity of data anda large sequence space make this task challenging, and to date only modelslimited to a small set of epitopes have achieved good performance. Here, weestablish a k-nearest-neighbor (K-NN) classifier as a strong baseline and thenpropose TITAN (Tcr epITope bimodal Attention Networks), a bimodal neuralnetwork that explicitly encodes both TCR sequences and epitopes to enable theindependent study of generalization capabilities to unseen TCRs and/orepitopes. Results: By encoding epitopes at the atomic level with SMILESsequences, we leverage transfer learning and data augmentation to enrich theinput data space and boost performance. TITAN achieves high performance in theprediction of specificity of unseen TCRs (ROC-AUC 0.87 in 10-fold CV) andsurpasses the results of the current state-of-the-art (ImRex) by a largemargin. Notably, our Levenshtein-distance-based K-NN classifier also exhibitscompetitive performance on unseen TCRs. While the generalization to unseenepitopes remains challenging, we report two major breakthroughs. First, bydissecting the attention heatmaps, we demonstrate that the sparsity ofavailable epitope data favors an implicit treatment of epitopes as classes.This may be a general problem that limits unseen epitope performance forsufficiently complex models. Second, we show that TITAN nevertheless exhibitssignificantly improved performance on unseen epitopes and is capable offocusing attention on chemically meaningful molecular structures.

TITAN: T Cell Receptor Specificity Prediction with Bimodal Attention Networks