Attention network for predicting T cell receptor-peptide binding can associate attention with interpretable protein structural properties

Koyama, Kyohei; Hashimoto, Kosuke; Nagao, Chioko; Mizuguchi, Kenji

doi:10.1101/2023.02.16.528799

Cited by 3 publications

(4 citation statements)

References 53 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The CDR3 region interacts with the amino acid residues on the peptide through various non-covalent interactions, such as hydrogen bonding, ionic interactions, and van der Waals forces, to establish stable binding. The three-dimensional (3D) structure, determined by these amino acid sequences, underlies the speci c binding of CDR3 to different epitopes 12,13 . Thus, individual sequence data alone are insu cient to capture all the information that governs CDR-epitope binding.…”

Section: Introductionmentioning

confidence: 99%

Predicting TCR sequences for unseen antigen epitopes using structural and sequence features

Zhang,

Ji,

Zhang

et al. 2024

Preprint

View full text Add to dashboard Cite

T-cell receptor (TCR) recognition of antigens is fundamental to the adaptive immune response. With the expansion of experimental techniques, a substantial database of matched TCR-antigen pairs has emerged, presenting opportunities for computational prediction models. However, the accurate forecasting of binding affinities for unseen antigen-TCR pairs remains a major challenge. Here, we present Convolutional-Self-Attention TCR (CATCR), a novel framework tailored to enhance the prediction of epitope and TCR interactions. Our approach integrates an encoder that concurrently processes structural and sequential data, utilizing convolutional neural networks (CNNs) to extract peptide features from residue contact matrices, as generated by OpenFold, and a Transformer to encode segment-based coded sequence. We further introduce CATCR-D, a discriminator equipped to assess binding by analyzing structural and sequence features of epitopes and CDR3-β regions. Additionally, the framework comprises CATCR-G, a generative module designed for CDR3-β sequences, which applies the pretrained encoder to deduce epitope characteristics and a Transformer decoder for predicting matching CDR3-β sequences. CATCR-D has shown exemplary feature extraction and generalization, achieving an AUROC of 0.89 on previously unseen epitope-TCR pairs and outperforming four benchmark models by a margin of 17.4%. CATCR-G has demonstrated high precision, recall, and F1 scores, surpassing 95% in BERT-score assessments. Our results indicate that CATCR is an effective tool for the prediction of unseen epitope-TCR interactions, and that incorporating structural insights significantly enhances our understanding of the general rules governing TCR-epitope recognition. The prediction of TCRs for novel epitopes using structural and sequence information is promising, and broadening the repository of experimental TCR-epitope data stands to further improve the precision of epitope-TCR binding predictions.

show abstract

Section: Introductionmentioning

confidence: 99%

Predicting TCR sequences for unseen antigen epitopes using structural and sequence features

Zhang,

Ji,

Zhang

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…More complex models [5,6,7,8] are also proposed for the classification task. Many deep learning models (NetTCR [9], DeepTCR [7], ImRex [8], tcrpred [10]) rely on convolutional neural networks to learn the TCR and peptide patterns in each interaction. TITAN [11], ATM-TCR [12], and tcrpred [10] further evaluate the pairwise interactions by crossing the TCR and peptide patterns in an attention structure.…”

Section: Introductionmentioning

confidence: 99%

“…Many deep learning models (NetTCR [9], DeepTCR [7], ImRex [8], tcrpred [10]) rely on convolutional neural networks to learn the TCR and peptide patterns in each interaction. TITAN [11], ATM-TCR [12], and tcrpred [10] further evaluate the pairwise interactions by crossing the TCR and peptide patterns in an attention structure. Some other tools, particularly ERGO-I [13] and pMTnet [14], use long short-term memory (LSTM) to learn the sequential information of TCR and peptides, and autoencoder layers to simultaneously improve the data understanding and reduce the feature space.…”

Section: Introductionmentioning

confidence: 99%

epiTCR-KDA: Knowledge Distillation model on Dihedral Angles for TCR-peptide prediction

Pham,

Su,

Nguyen

et al. 2024

Preprint

View full text Add to dashboard Cite

Motivation: Antigen recognition by T-cell receptors (TCRs) triggers cascades of immune responses. Successful predictions of the TCR and antigen (as peptide) bindings therefore signify the advancements in immunotherapy. However, most of current TCR-peptide interaction predictors fail to predict unseen data. This limitation may be derived from the conventional usage of TCR and/or peptide sequences as input, which may not adequately reflect their structural characteristics. Therefore, incorporating the TCR and peptide structural information into the prediction model to improve the generalizability is necessary. Results: We presented epiTCR-KDA as a new predictor of TCR-peptide binding that utilises structural information, specifically the dihedral angles between the residues of both the peptide and the TCR. This structural descriptor was integrated into a model constructed using knowledge distillation to enhance its generalizability. The epiTCR-KDA demonstrated competitive prediction performance, with an AUC of 0.99 for seen data and AUC of 0.86 for unseen data. Across multiple public datasets, epiTCR-KDA consistently outperformed other predictors, such as epiTCR, NetTCR, BERTrand, TEIM-Seq, TEINet, and ImRex, maintaining a median AUC of 0.9 (ranging from 0.82 to 0.91). Further analysis of epiTCR-KDA performance indicated that the cosine similarity of the dihedral angle vectors between the unseen testing data and training data is crucial for its stable performance. In conclusion, our epiTCR-KDA model, with its capacity to predict for unseen data, has brought us one step closer toward the development of a highly effective pipeline for affordable antigen-based immunotherapy. Availability and implementation: epiTCR-KDA is available on GitHub (https://github.com/ddiem-ri-4D/epiTCR-KDA) .

show abstract

“…The CDR3 region interacts with the amino acid residues on the peptide through various non-covalent interactions, such as hydrogen bonding, ionic interactions and van der Waals forces, to establish stable binding. The three-dimensional (3D) structure, determined by these amino acid sequences, underlies the specific binding of CDR3 to different epitopes [ 14 , 15 ]. Thus, individual sequence data alone cannot capture all the information governing CDR–epitope binding.…”

Section: Introductionmentioning

confidence: 99%

Predicting TCR sequences for unseen antigen epitopes using structural and sequence features

Ji,

Wang,

Zhang

et al. 2024

Briefings in Bioinformatics

View full text Add to dashboard Cite

T-cell receptor (TCR) recognition of antigens is fundamental to the adaptive immune response. With the expansion of experimental techniques, a substantial database of matched TCR–antigen pairs has emerged, presenting opportunities for computational prediction models. However, accurately forecasting the binding affinities of unseen antigen–TCR pairs remains a major challenge. Here, we present convolutional-self-attention TCR (CATCR), a novel framework tailored to enhance the prediction of epitope and TCR interactions. Our approach utilizes convolutional neural networks to extract peptide features from residue contact matrices, as generated by OpenFold, and a transformer to encode segment-based coded sequences. We introduce CATCR-D, a discriminator that can assess binding by analyzing the structural and sequence features of epitopes and CDR3-β regions. Additionally, the framework comprises CATCR-G, a generative module designed for CDR3-β sequences, which applies the pretrained encoder to deduce epitope characteristics and a transformer decoder for predicting matching CDR3-β sequences. CATCR-D achieved an AUROC of 0.89 on previously unseen epitope–TCR pairs and outperformed four benchmark models by a margin of 17.4%. CATCR-G has demonstrated high precision, recall and F1 scores, surpassing 95% in bidirectional encoder representations from transformers score assessments. Our results indicate that CATCR is an effective tool for predicting unseen epitope–TCR interactions. Incorporating structural insights enhances our understanding of the general rules governing TCR–epitope recognition significantly. The ability to predict TCRs for novel epitopes using structural and sequence information is promising, and broadening the repository of experimental TCR–epitope data could further improve the precision of epitope–TCR binding predictions.

show abstract

Attention network for predicting T cell receptor-peptide binding can associate attention with interpretable protein structural properties

Cited by 3 publications

References 53 publications

Predicting TCR sequences for unseen antigen epitopes using structural and sequence features

Predicting TCR sequences for unseen antigen epitopes using structural and sequence features

epiTCR-KDA: Knowledge Distillation model on Dihedral Angles for TCR-peptide prediction

Predicting TCR sequences for unseen antigen epitopes using structural and sequence features

Contact Info

Product

Resources

About