To construct high-quality benchmark datasets, all the TCR-pMHC pairs used in this study were mainly collected from the databases VDJdb and IEDB as well as TetTCR-seq data.
Training Dataset :
Peptide-CDR3α:
Download File
Peptide-CDR3β:
Download File
Testing Dataset :
Peptide-CDR3α:
Download File
Peptide-CDR3β:
Download File
Independent Testing Dataset I (COVID-19 dataset) :
Peptide-CDR3α:
Download File
Peptide-CDR3β:
Download File
Independent Testing Dataset II (IEDB dataset) :
Peptide-CDR3αβ:
Download File.
Independent Testing Dataset III (using for performance comparison) :
McPAS-shared:
Download File
McPAS-unique:
Download File
COVID-19:
Download File
A remainder of peptide-TCR pairs in VDJdb were filtered out during benchmark dataset construction to assess the prediction ability of ensemble classifiers for TCR cross-reactivity.
Cross-reactivity Dataset :
Peptide-CDR3α:
Download File
Peptide-CDR3β:
Download File
Supplementary Data ：
Supplementary Data 1:
Download File
Supplementary Data 2:
Download File
Supplementary Data 3:
Download File
Supplementary Data 4:
Download File
Supplementary Data 5:
Download File
Supplementary Data 6:
Download File
Supplementary Data 7:
Download File
Source Code ：
Hosted on GitHub