TCRfp: a new fingerprint-based approach for TCR repertoire analysis

Francesca Mayol-Rullan,Marine Bugnon,Marta A. S. Perez,Vincent Zoete

DOI: https://doi.org/10.1101/2023.12.19.572261

2024-02-01

Abstract:The development of cancer immunotherapy has accelerated in recent years. Understanding the specificity of T cell receptors (TCR) for peptides presented by the major histocompatibility complex (pMHC) is a major step towards improving immunotherapy approaches, such as adoptive cell transfer and peptide vaccination. Despite recent computational advances, the unambiguous pairing of TCR with pMHC, from pools of thousands of candidates, remains out of reach. To tackle this challenge, we have developed a new tool that converts the 3D structure of TCR into individual one-dimensional structural fingerprints (TCRfp). We have modelled over 10’000 3D structures of paired TCR alpha and beta chains with known sequences and pMHC specificity and encoded them into 1D TCRfp. For future clinical needs, we have translated the TCR modelling process into a fast pipeline. Similarity measures between TCR FPs correlate with their ability to recognise similar or identical epitopes in the training set and in the external validation sets. TCRfp constitutes the first rapid approach for high-throughput TCR comparison and repertoire analysis based on molecular 3D structures, which is efficient enough to complement sequence-based approaches.

Cancer Biology

What problem does this paper attempt to address?

### What problem does this paper attempt to solve? This paper aims to solve the problem of specific recognition between T - cell receptors (TCR) and peptides presented by major histocompatibility complexes (pMHC). Specifically, the authors attempt to develop a new method to accurately predict the ability of TCR to recognize specific pMHC, which is crucial for improving cancer immunotherapies such as adoptive cell transfer and peptide vaccination. Currently, despite the progress in computing technology, it is still a challenge to clearly pair TCR with pMHC from thousands of candidates. To this end, the authors have developed a new 3D - structure - based tool - **TCRfp**, which converts the 3D structure of TCR into a one - dimensional structural fingerprint, enabling rapid and efficient TCR comparison and library analysis. ### Main objectives 1. **Improve the accuracy of TCR - pMHC specificity prediction**: By converting the 3D structure of TCR into a one - dimensional fingerprint (TCRfp), the authors hope to quickly identify TCRs with similar or identical epitope - recognition abilities in large - scale datasets. 2. **Develop a fast and scalable TCR modeling and analysis pipeline**: The authors have designed an automated pipeline that can generate high - quality TCR 3D models from sequences and use these models to calculate TCRfp, enabling high - throughput TCR analysis. 3. **Explore methods to optimize fingerprint definitions**: Systematically explore different fingerprint definitions through genetic algorithms (GA) to find the best similarity estimation method and ensure that the fingerprints can better reflect the binding specificity of TCR. ### Key innovation points - **TCRfp**: Encode the 3D structure of TCR into a one - dimensional fingerprint, and use the ES5D algorithm to extend the classical shape - matching method by adding charge and hydrophobicity dimensions. - **Fast modeling and fingerprint calculation**: Generate TCR 3D models through an automated pipeline and calculate TCRfp within a few minutes, greatly improving the processing speed. - **Structure - based similarity measurement**: Measure the similarity between TCRs by calculating the Manhattan distance between fingerprints, and then infer whether they can recognize the same pMHC. In conclusion, this paper provides a brand - new, structure - based and efficient method to analyze and compare TCRs by introducing TCRfp, which is expected to significantly improve the accuracy and efficiency of TCR - pMHC specificity prediction.

TCRfp: a new fingerprint-based approach for TCR repertoire analysis

Predicting Antigen‐Specificities of Orphan T Cell Receptors from Cancer Patients with TCRpcDist

Abstract 3559: A validated bioinformatics tool-set for predicting TCR specificity

A Computational Strategy for the Rapid Identification and Ranking of Patient‐Specific T cell Receptors Bound to Neoantigens

[ACNP therapy for DCMP-resistant acute leukemia (author's transl)].

Integrating T-cell Receptor and Transcriptome for Large-Scale Single-Cell Immune Profiling Analysis

TCR-L: an analysis tool for evaluating the association between the T-cell receptor repertoire and clinical phenotypes

SwiftTCR: Efficient Computational Docking protocol of TCRpMHC-I Complexes Using Restricted Rotation Matrices

Rapid parallel reconstruction and specificity screening of hundreds of T cell receptors

TCRpred: incorporating T-cell receptor repertoire for clinical outcome prediction

TCR meta-clonotypes for biomarker discovery with tcrdist3 enabled identification of public, HLA-restricted clusters of SARS-CoV-2 TCRs

T-Cell Receptor Cognate Target Prediction Based on Paired α and β Chain Sequence and Structural CDR Loop Similarities

FUME-TCRseq enables sensitive and accurate sequencing of the T-cell receptor from limited input of degraded RNA

TCRfinder: Improved TCR virtual screening for novel antigenic peptides with tailored language models

NetTCR-2.1: Lessons and guidance on how to develop models for TCR specificity predictions

Analysis of the Repertoire Features of TCR Beta Chain CDR3 in Human by High-Throughput Sequencing

TCR-TRANSLATE: Conditional Generation of Real Antigen Specific T-cell Receptor Sequences

Rigorous benchmarking of T-cell receptor repertoire profiling methods for cancer RNA sequencing

tcrLM: a lightweight protein language model for predicting T cell receptor and epitope binding specificity

Deciphering CD4+ T cell specificity using novel MHC–TCR chimeric receptors

Identification of clinically relevant T cell receptors for personalized T cell therapy using combinatorial algorithms