Reconstituting T cell receptor selection in-silico.

Area of Science:

• Immunology
• Computational Biology
• Genetics

Background:

• T cell receptors (TCRs) are crucial for adaptive immunity, recognizing antigens presented by major histocompatibility complexes (MHCs).
• T cell selection is a critical process that eliminates T cells with non-functional or self-reactive TCRs, preventing autoimmunity.
• The specific features of TCR protein sequences that dictate selection outcomes remain largely unknown.

Purpose of the Study:

• To develop and validate a computational approach for identifying patterns in TCR protein sequences associated with T cell selection.
• To differentiate between T cell receptors (TCRs) before and after the selection process using machine learning.

Main Methods:

• Utilized T cell receptor (TCR) gene sequencing to obtain sequence data from developing and mature T cells.
• Applied machine learning algorithms to analyze TCR protein sequences and identify predictive patterns related to selection.
• Validated the trained machine learning models by classifying TCRs from distinct developmental stages.

Main Results:

• Successfully developed machine learning models capable of classifying TCRs based on their presence before or after T cell selection.
• Identified specific patterns within TCR protein sequences that correlate with the outcome of T cell selection.
• Demonstrated the efficacy of the computational approach in distinguishing selected from unselected T cell receptors.

Conclusions:

• The developed computational method provides a novel way to investigate the molecular basis of T cell selection.
• Understanding TCR sequence determinants of selection can offer insights into the development of autoimmune diseases.
• This approach may facilitate future research into immune system regulation and therapeutic strategies for immune-related disorders.