Machine learning analysis of the T cell receptor repertoire identifies sequence features of self-reactivity

Johannes Textor; Franka Buytenhuijs; Dakota Rogers; Ève Mallet Gauthier; Shabaz Sultan; Inge M.N. Wortel; Kathrin Kalies; Anke Fähnrich; René Pagel; Heather J. Melichar; Jürgen Westermann; Judith N. Mandl

doi:10.1016/j.cels.2023.11.004

Machine learning analysis of the T cell receptor repertoire identifies sequence features of self-reactivity

Johannes Textor^*, Franka Buytenhuijs, Dakota Rogers, Ève Mallet Gauthier, Shabaz Sultan, Inge M.N. Wortel, Kathrin Kalies, Anke Fähnrich, René Pagel, Heather J. Melichar, Jürgen Westermann, Judith N. Mandl^*

^*Korrespondierende/r Autor/-in für diese Arbeit

11 Zitate (Scopus)

Abstract

The T cell receptor (TCR) determines specificity and affinity for both foreign and self-peptides presented by the major histocompatibility complex (MHC). Although the strength of TCR interactions with self-pMHC impacts T cell function, it has been challenging to identify TCR sequence features that predict T cell fate. To discern patterns distinguishing TCRs from naive CD4⁺ T cells with low versus high self-reactivity, we used data from 42 mice to train a machine learning (ML) algorithm that identifies population-level differences between TCRβ sequence sets. This approach revealed that weakly self-reactive T cell populations were enriched for longer CDR3β regions and acidic amino acids. We tested our ML predictions of self-reactivity using retrogenic mice with fixed TCRβ sequences. Extrapolating our analyses to independent datasets, we predicted high self-reactivity for regulatory T cells and slightly reduced self-reactivity for T cells responding to chronic infections. Our analyses suggest a potential trade-off between TCR repertoire diversity and self-reactivity. A record of this paper's transparent peer review process is included in the supplemental information.

Originalsprache	Englisch
Zeitschrift	Cell Systems
Jahrgang	14
Ausgabenummer	12
Seiten (von - bis)	1059-1073.e5
ISSN	2405-4712
DOIs	https://doi.org/10.1016/j.cels.2023.11.004
Publikationsstatus	Veröffentlicht - 20.12.2023

Fördermittel

We would like to thank P. D’Arcy, G. Perreault, and animal facility staff at McGill for their excellent care of our animal colony; Ann Feeney (Scripps) for sharing TdT −/− mice; Thierry Mallevaey for providing the TCRβ constructs used for the retrogenic experiments; Julien Leconte and Camille Stegen at the McGill University Cell Vision Core Facility for performing all cell sorts; and P. Artusa for contributions to T cell sorting. D.R. was funded by a Frederick Banting and Charles Best Canada Graduate Scholarships doctoral award ( CIHR CGS-D ) and a Tomlinson Doctoral Fellowship (McGill University). È.M.G. is the recipient of a Natural Sciences and Engineering Research Council of Canada Graduate Scholarship doctoral award ( CGS D ). H.J.M. is a Fonds de recherche du Quebec-Sante Senior Research Scholar. J.N.M. holds a Canada Research Chair for immune cell dynamics. This research was supported by an NSERC discovery grant ( 2016-03808 ) and a McGill start-up fund to J.N.M., a CIHR ( PJT-168862 ) grant jointly to J.N.M. and H.J.M., and an NWO Vidi grant ( VI.Vidi.192.084 ) to J.T. K.K. was funded by the German Research Foundation (DFG) within the framework of the Schleswig-Holstein Excellence Cluster I&I (EXC 306, Inflammation at Interfaces, project XTP4). J.W. and K.K. were funded by the graduate schools GRK 1727/2 and the TR-SFB654 project C4 at the University of Lübeck . We would like to thank P. D'Arcy, G. Perreault, and animal facility staff at McGill for their excellent care of our animal colony; Ann Feeney (Scripps) for sharing TdT−/− mice; Thierry Mallevaey for providing the TCRβ constructs used for the retrogenic experiments; Julien Leconte and Camille Stegen at the McGill University Cell Vision Core Facility for performing all cell sorts; and P. Artusa for contributions to T cell sorting. D.R. was funded by a Frederick Banting and Charles Best Canada Graduate Scholarships doctoral award (CIHR CGS-D) and a Tomlinson Doctoral Fellowship (McGill University). È.M.G. is the recipient of a Natural Sciences and Engineering Research Council of Canada Graduate Scholarship doctoral award (CGS D). H.J.M. is a Fonds de recherche du Quebec-Sante Senior Research Scholar. J.N.M. holds a Canada Research Chair for immune cell dynamics. This research was supported by an NSERC discovery grant (2016-03808) and a McGill start-up fund to J.N.M. a CIHR (PJT-168862) grant jointly to J.N.M. and H.J.M. and an NWO Vidi grant (VI.Vidi.192.084) to J.T. K.K. was funded by the German Research Foundation (DFG) within the framework of the Schleswig-Holstein Excellence Cluster I&I (EXC 306, Inflammation at Interfaces, project XTP4). J.W. and K.K. were funded by the graduate schools GRK 1727/2 and the TR-SFB654 project C4 at the University of Lübeck. J.N.M. and J.T. conceived, designed, and supervised the project and research, with key input from H.J.M. and I.M.N.W. F.B. and J.T. analyzed the TCR-seq data. F.B. developed the ML algorithm with input from S.S. under the supervision of J.T. K.K. A.F. and R.P. performed the TCR-seq on cells sorted by D.R. and J.N.M. under the supervision of J.W. È.M.G. performed the TCR retrogenic mouse experiments designed and supervised by H.J.M. The manuscript was written by J.N.M. J.T. D.R. and F.B. with feedback from all other authors. The authors declare no competing interests.

UN SDGs

Dieser Output leistet einen Beitrag zu folgendem(n) Ziel(en) für nachhaltige Entwicklung

SDG 3 – Gesundheit und Wohlergehen

Strategische Forschungsbereiche und Zentren

Forschungsschwerpunkt: Infektion und Entzündung - Zentrum für Infektions- und Entzündungsforschung Lübeck (ZIEL)

DFG-Fachsystematik

2.21-05 Immunologie
2.22-04 Anatomie und Physiologie

Dokumente

10.1016/j.cels.2023.11.004

Weitere Links

Link to publication in Scopus

Zitieren

Textor, J., Buytenhuijs, F., Rogers, D., Gauthier, È. M., Sultan, S., Wortel, I. M. N., Kalies, K., Fähnrich, A., Pagel, R., Melichar, H. J., Westermann, J., & Mandl, J. N. (2023). Machine learning analysis of the T cell receptor repertoire identifies sequence features of self-reactivity. Cell Systems, 14(12), 1059-1073.e5. https://doi.org/10.1016/j.cels.2023.11.004

@article{864587fd42f143daae75fb1a9d3dfc47,

title = "Machine learning analysis of the T cell receptor repertoire identifies sequence features of self-reactivity",

abstract = "The T cell receptor (TCR) determines specificity and affinity for both foreign and self-peptides presented by the major histocompatibility complex (MHC). Although the strength of TCR interactions with self-pMHC impacts T cell function, it has been challenging to identify TCR sequence features that predict T cell fate. To discern patterns distinguishing TCRs from naive CD4+ T cells with low versus high self-reactivity, we used data from 42 mice to train a machine learning (ML) algorithm that identifies population-level differences between TCRβ sequence sets. This approach revealed that weakly self-reactive T cell populations were enriched for longer CDR3β regions and acidic amino acids. We tested our ML predictions of self-reactivity using retrogenic mice with fixed TCRβ sequences. Extrapolating our analyses to independent datasets, we predicted high self-reactivity for regulatory T cells and slightly reduced self-reactivity for T cells responding to chronic infections. Our analyses suggest a potential trade-off between TCR repertoire diversity and self-reactivity. A record of this paper's transparent peer review process is included in the supplemental information.",

author = "Johannes Textor and Franka Buytenhuijs and Dakota Rogers and Gauthier, \{{\`E}ve Mallet\} and Shabaz Sultan and Wortel, \{Inge M.N.\} and Kathrin Kalies and Anke F{\"a}hnrich and Ren{\'e} Pagel and Melichar, \{Heather J.\} and J{\"u}rgen Westermann and Mandl, \{Judith N.\}",

note = "Publisher Copyright: {\textcopyright} 2023 The Author(s)",

year = "2023",

month = dec,

day = "20",

doi = "10.1016/j.cels.2023.11.004",

language = "English",

volume = "14",

pages = "1059--1073.e5",

journal = "Cell Systems",

issn = "2405-4712",

publisher = "Cell Press",

number = "12",

}

TY - JOUR

T1 - Machine learning analysis of the T cell receptor repertoire identifies sequence features of self-reactivity

AU - Textor, Johannes

AU - Buytenhuijs, Franka

AU - Rogers, Dakota

AU - Gauthier, Ève Mallet

AU - Sultan, Shabaz

AU - Wortel, Inge M.N.

AU - Kalies, Kathrin

AU - Fähnrich, Anke

AU - Pagel, René

AU - Melichar, Heather J.

AU - Westermann, Jürgen

AU - Mandl, Judith N.

PY - 2023/12/20

Y1 - 2023/12/20

N2 - The T cell receptor (TCR) determines specificity and affinity for both foreign and self-peptides presented by the major histocompatibility complex (MHC). Although the strength of TCR interactions with self-pMHC impacts T cell function, it has been challenging to identify TCR sequence features that predict T cell fate. To discern patterns distinguishing TCRs from naive CD4+ T cells with low versus high self-reactivity, we used data from 42 mice to train a machine learning (ML) algorithm that identifies population-level differences between TCRβ sequence sets. This approach revealed that weakly self-reactive T cell populations were enriched for longer CDR3β regions and acidic amino acids. We tested our ML predictions of self-reactivity using retrogenic mice with fixed TCRβ sequences. Extrapolating our analyses to independent datasets, we predicted high self-reactivity for regulatory T cells and slightly reduced self-reactivity for T cells responding to chronic infections. Our analyses suggest a potential trade-off between TCR repertoire diversity and self-reactivity. A record of this paper's transparent peer review process is included in the supplemental information.

AB - The T cell receptor (TCR) determines specificity and affinity for both foreign and self-peptides presented by the major histocompatibility complex (MHC). Although the strength of TCR interactions with self-pMHC impacts T cell function, it has been challenging to identify TCR sequence features that predict T cell fate. To discern patterns distinguishing TCRs from naive CD4+ T cells with low versus high self-reactivity, we used data from 42 mice to train a machine learning (ML) algorithm that identifies population-level differences between TCRβ sequence sets. This approach revealed that weakly self-reactive T cell populations were enriched for longer CDR3β regions and acidic amino acids. We tested our ML predictions of self-reactivity using retrogenic mice with fixed TCRβ sequences. Extrapolating our analyses to independent datasets, we predicted high self-reactivity for regulatory T cells and slightly reduced self-reactivity for T cells responding to chronic infections. Our analyses suggest a potential trade-off between TCR repertoire diversity and self-reactivity. A record of this paper's transparent peer review process is included in the supplemental information.

UR - https://www.scopus.com/pages/publications/85180310325

U2 - 10.1016/j.cels.2023.11.004

DO - 10.1016/j.cels.2023.11.004

M3 - Journal articles

C2 - 38061355

AN - SCOPUS:85180310325

SN - 2405-4712

VL - 14

SP - 1059-1073.e5

JO - Cell Systems

JF - Cell Systems

IS - 12

ER -

Machine learning analysis of the T cell receptor repertoire identifies sequence features of self-reactivity

Abstract

Fördermittel

UN SDGs

Strategische Forschungsbereiche und Zentren

DFG-Fachsystematik

Dokumente

Weitere Links

Fingerprint

Zitieren