Deep learning-based prediction of the T cell receptor–antigen binding specificity

Lu, Tianshi; Zhang, Ze; Zhu, James; Wang, Yunguan; Jiang, Peixin; Xiao, Xue; Bernatchez, Chantale; Heymach, John V.; Gibbons, Don L.; Wang, Jun; Xu, Lin; Reuben, Alexandre; Wang, Tao

doi:10.1038/s42256-021-00383-2

Cited by 148 publications

(204 citation statements)

References 78 publications

Supporting

Mentioning

204

Contrasting

Order By: Relevance

“…As a proof of concept, we apply our pipeline to two recently published deep-learning models for TCR-binding prediction, ( Weber et al , 2021 ) and pMHC-TCR-binding prediction network ( Lu et al , 2021 ). To keep our proof of concept simple, we analyze both models only on a single epitope : the peptide KLGGALQAK from Cytomegalovirus.…”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

DECODE: a computational pipeline to discover T cell receptor binding rules

et al. 2022

View full text Add to dashboard Cite

Motivation Understanding the mechanisms underlying T cell receptor (TCR) binding is of fundamental importance to understanding adaptive immune responses. A better understanding of the biochemical rules governing TCR binding can be used, e.g. to guide the design of more powerful and safer T cell-based therapies. Advances in repertoire sequencing technologies have made available millions of TCR sequences. Data abundance has, in turn, fueled the development of many computational models to predict the binding properties of TCRs from their sequences. Unfortunately, while many of these works have made great strides toward predicting TCR specificity using machine learning, the black-box nature of these models has resulted in a limited understanding of the rules that govern the binding of a TCR and an epitope. Results We present an easy-to-use and customizable computational pipeline, DECODE, to extract the binding rules from any black-box model designed to predict the TCR-epitope binding. DECODE offers a range of analytical and visualization tools to guide the user in the extraction of such rules. We demonstrate our pipeline on a recently published TCR-binding prediction model, TITAN, and show how to use the provided metrics to assess the quality of the computed rules. In conclusion, DECODE can lead to a better understanding of the sequence motifs that underlie TCR binding. Our pipeline can facilitate the investigation of current immunotherapeutic challenges, such as cross-reactive events due to off-target TCR binding. Availability and implementation Code is available publicly at https://github.com/phineasng/DECODE. Supplementary information Supplementary data are available at Bioinformatics online.

show abstract

Section: Resultsmentioning

confidence: 99%

“…( Lu et al , 2021 ) is a transfer learning-based model that predicts TCR -epitope binding based on an Atchley factor encoding of the CDR3 loop of the TCR and a joint encoding of epitope and MHC similar to the netMHCpan model. outputs the fractional rank of each TCR compared to 10 000 background receptors, i.e.…”

Section: Resultsmentioning

confidence: 99%

DECODE: a computational pipeline to discover T cell receptor binding rules

et al. 2022

View full text Add to dashboard Cite

show abstract

“…Considering the diversity of potential antigens (and the complexity of MHC), developing models to predict cognate pMHCs of a given TCR is challenging. Recent efforts have reframed this problem to predict whether a TCR will recognize a given peptide and have shown some success with [107] and without [108,109] modeling MHC restriction. Several challenges in this endeavor, however, remain unsolved.…”

Section: Box 1 Limitations Of the Barcode Methodsmentioning

confidence: 99%

TCR-sequencing in cancer and autoimmunity: barcodes and beyond

Pauken¹,

Lagattuta²,

Lu³

et al. 2022

Trends in Immunology

View full text Add to dashboard Cite

“…And the other is to dynamically calculate the average of the performance scores for the user-chosen metrics and to rank the chosen DR methods based on the average scores (top row). This web tool is hosted on the Database for Actionable Immunology (DBAI) website (https://dbai.biohpc.swmed.edu/) (31,32,40).…”

Section: Development Of the Cytof Dr Playground Webservermentioning

confidence: 99%

Comparative Analysis of Dimension Reduction Methods for Cytometry by Time-of-Flight Data

Wang

Yang

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Recently developed single-cell profiling technologies, such as single cell sequencing (scRNA-seq) and mass cytometry (CyTOF), provide critical biological insights from a large number of cells and a large number of features. While experimental and informatic techniques around scRNA-seq are much more advanced, research around CyTOF data analysis has severely lagged behind. However, CyTOF profiles the proteomics makeup of the single cells and are more closely related to disease phenotypes than scRNA-seq. CyTOF data are also dramatically different from scRNA-seq data in many aspects. This calls for the evaluation and development of statistical methods for CyTOF data. Dimension reduction (DR) is one of the most critical first steps of analyses of scRNA-seq and CyTOF data. Here, we benchmark 20 of these methods on 110 real and 425 synthetic CyTOF datasets, including 10 Imaging CyTOF datasets, for accuracy, scalability and usability. In particular, we checked the concordance of DR for CyTOF data against scRNA-seq data that were generated from the same samples. We found that SAUCIE and scvis surprisingly perform well across all categories, whereas UMAP particularly excels in downstream analyses and MDS preserves local and global structure well. Our results highlight the complementarity of existing tools, and that the choice of method should depend on the underlying data structure of the single cells. Based on these results, we develop a set of guidelines to help users select the best DR method for their dataset. Our freely available data and evaluation pipeline will aid in the development of improved DR algorithms specifically tailored to analyze data generated from the increasingly popular CyTOF technique.

show abstract

Deep learning-based prediction of the T cell receptor–antigen binding specificity

Cited by 148 publications

References 78 publications

DECODE: a computational pipeline to discover T cell receptor binding rules

DECODE: a computational pipeline to discover T cell receptor binding rules

TCR-sequencing in cancer and autoimmunity: barcodes and beyond

Comparative Analysis of Dimension Reduction Methods for Cytometry by Time-of-Flight Data

Contact Info

Product

Resources

About