/TCRs-pMHC

Datasets containing CDR3 Beta sequences and peptide sequences, with a true label indication whether the combination binds or not.

TCRs-pMHC

This repository share the dataversion used.

Dataversions

  • dbase,strict
  • dbase,uniform
  • dbal
  • dimbal

All dataversion contain negative and positive examples. All dataversion has 5 folds each contain a training and test set.

*d tpp_dataset.csv.gz is the TPP-dataset not splittet apart with additional information like referencelink and MHC. to unzip use gunzip cmd

Citation

Use for this benchmark paper following citation

@article{weber2021titan
    author = {Deng, Lihua and Ly, Cedric and Abdollahi, Sina and Zhao, Yu and Prinz, Immo and Bonn, Stefan},
    title = "{Performance comparison of TCR-pMHC prediction tools reveals a strong data dependency}",
    journal = {Frontiers in Immunology},
    volume = {14},
    number = {},
    pages = {},
    year = {2023},
    month = {},
    issn = {1664-3224},
    doi = {10.3389/fimmu.2023.1128326},
    url = {https://www.frontiersin.org/articles/10.3389/fimmu.2023.1128326}
}