TCRex: a web tool for the prediction of TCR–epitope recognition
Welcome to TCRex!
TCRex predicts TCR–epitope binding for human T cell receptors (TCRs) using TCR beta chain information, i.e. the CDR3 amino acid sequence and the corresponding V/J genes. It is based on random forest classifiers trained on epitope-specific TCR data collected from the manually curated catalogue of pathology-associated T cell receptor sequences (McPAS-TCR), the VDJ database (VDJdb) and the ImmuneCODETM database. In total prediction models for 100 different epitopes, consisting of 93 viral and 7 cancer epitopes, are provided. Check the statistics page for detailed information about the performance of the individual prediction models.
Although TCRex supports a wide range of epitopes it might not include a prediction model for your epitope(s) of interest. However, in this case you can use our optimized machine learning and prediction workflow to train your own custom model using the new epitopes page.
In addition, TCRex can be used to perform epitope-specificity enrichment analyses to identify the epitopes that are targeted by the uploaded TCR data set.
Detailed information on how to use TCRex and interpret the results is provided on the instructions page.
Predict TCR–epitope binding
TCRex has been developed as a joint project between AUDACIS, biomina, and the Adrem Data Lab. More information about these collaborators can be found on the about page.
Please cite the following manuscript if you want use TCRex in your work:
S. Gielis, P. Moris, W. Bittremieux, N. De Neuter, B. Ogunjimi, K. Laukens and P. Meysman. Detection of Enriched T Cell Epitope Specificity in Full T Cell Receptor Sequence Repertoires. Frontiers in Immunology, 2019. (doi: 10.3389/fimmu.2019.02820)