Comparison experiments are conducted to evaluate different input embeddings using the proposed GTE model on four datasets with the StrictTCR setting; both AUC and AUPR scores are reported; each result derives from five-fold cross-validation and includes standard deviations; the best scores are marked in bold
| Embeddings . | TEINet Dataset . | pMTnet Dataset . | VDJdb Dataset . | McPAS Dataset . | ||||
|---|---|---|---|---|---|---|---|---|
| . | AUC . | AUPR . | AUC . | AUPR . | AUC . | AUPR . | AUC . | AUPR . |
| ESM2(650M) | 0.824 |$\pm $| 0.008 | 0.459 |$\pm $| 0.017 | 0.899 |$\pm $| 0.003 | 0.632 |$\pm $| 0.008 | 0.829 |$\pm $| 0.005 | 0.466 |$\pm $| 0.009 | 0.742 |$\pm $| 0.023 | 0.305 |$\pm $| 0.032 |
| ESM2(3B) | 0.814 |$\pm $| 0.015 | 0.441 |$\pm $| 0.029 | 0.897 |$\pm $| 0.003 | 0.625 |$\pm $| 0.007 | 0.827 |$\pm $| 0.005 | 0.462 |$\pm $| 0.009 | 0.746 |$\pm $| 0.019 | 0.308 |$\pm $| 0.030 |
| TCRpeg | 0.839|$\pm $| 0.005 | 0.486|$\pm $| 0.012 | 0.911|$\pm $| 0.003 | 0.655|$\pm $| 0.007 | 0.847|$\pm $| 0.005 | 0.500|$\pm $| 0.010 | 0.813|$\pm $| 0.014 | 0.390|$\pm $| 0.031 |
| Embeddings . | TEINet Dataset . | pMTnet Dataset . | VDJdb Dataset . | McPAS Dataset . | ||||
|---|---|---|---|---|---|---|---|---|
| . | AUC . | AUPR . | AUC . | AUPR . | AUC . | AUPR . | AUC . | AUPR . |
| ESM2(650M) | 0.824 |$\pm $| 0.008 | 0.459 |$\pm $| 0.017 | 0.899 |$\pm $| 0.003 | 0.632 |$\pm $| 0.008 | 0.829 |$\pm $| 0.005 | 0.466 |$\pm $| 0.009 | 0.742 |$\pm $| 0.023 | 0.305 |$\pm $| 0.032 |
| ESM2(3B) | 0.814 |$\pm $| 0.015 | 0.441 |$\pm $| 0.029 | 0.897 |$\pm $| 0.003 | 0.625 |$\pm $| 0.007 | 0.827 |$\pm $| 0.005 | 0.462 |$\pm $| 0.009 | 0.746 |$\pm $| 0.019 | 0.308 |$\pm $| 0.030 |
| TCRpeg | 0.839|$\pm $| 0.005 | 0.486|$\pm $| 0.012 | 0.911|$\pm $| 0.003 | 0.655|$\pm $| 0.007 | 0.847|$\pm $| 0.005 | 0.500|$\pm $| 0.010 | 0.813|$\pm $| 0.014 | 0.390|$\pm $| 0.031 |
Comparison experiments are conducted to evaluate different input embeddings using the proposed GTE model on four datasets with the StrictTCR setting; both AUC and AUPR scores are reported; each result derives from five-fold cross-validation and includes standard deviations; the best scores are marked in bold
| Embeddings . | TEINet Dataset . | pMTnet Dataset . | VDJdb Dataset . | McPAS Dataset . | ||||
|---|---|---|---|---|---|---|---|---|
| . | AUC . | AUPR . | AUC . | AUPR . | AUC . | AUPR . | AUC . | AUPR . |
| ESM2(650M) | 0.824 |$\pm $| 0.008 | 0.459 |$\pm $| 0.017 | 0.899 |$\pm $| 0.003 | 0.632 |$\pm $| 0.008 | 0.829 |$\pm $| 0.005 | 0.466 |$\pm $| 0.009 | 0.742 |$\pm $| 0.023 | 0.305 |$\pm $| 0.032 |
| ESM2(3B) | 0.814 |$\pm $| 0.015 | 0.441 |$\pm $| 0.029 | 0.897 |$\pm $| 0.003 | 0.625 |$\pm $| 0.007 | 0.827 |$\pm $| 0.005 | 0.462 |$\pm $| 0.009 | 0.746 |$\pm $| 0.019 | 0.308 |$\pm $| 0.030 |
| TCRpeg | 0.839|$\pm $| 0.005 | 0.486|$\pm $| 0.012 | 0.911|$\pm $| 0.003 | 0.655|$\pm $| 0.007 | 0.847|$\pm $| 0.005 | 0.500|$\pm $| 0.010 | 0.813|$\pm $| 0.014 | 0.390|$\pm $| 0.031 |
| Embeddings . | TEINet Dataset . | pMTnet Dataset . | VDJdb Dataset . | McPAS Dataset . | ||||
|---|---|---|---|---|---|---|---|---|
| . | AUC . | AUPR . | AUC . | AUPR . | AUC . | AUPR . | AUC . | AUPR . |
| ESM2(650M) | 0.824 |$\pm $| 0.008 | 0.459 |$\pm $| 0.017 | 0.899 |$\pm $| 0.003 | 0.632 |$\pm $| 0.008 | 0.829 |$\pm $| 0.005 | 0.466 |$\pm $| 0.009 | 0.742 |$\pm $| 0.023 | 0.305 |$\pm $| 0.032 |
| ESM2(3B) | 0.814 |$\pm $| 0.015 | 0.441 |$\pm $| 0.029 | 0.897 |$\pm $| 0.003 | 0.625 |$\pm $| 0.007 | 0.827 |$\pm $| 0.005 | 0.462 |$\pm $| 0.009 | 0.746 |$\pm $| 0.019 | 0.308 |$\pm $| 0.030 |
| TCRpeg | 0.839|$\pm $| 0.005 | 0.486|$\pm $| 0.012 | 0.911|$\pm $| 0.003 | 0.655|$\pm $| 0.007 | 0.847|$\pm $| 0.005 | 0.500|$\pm $| 0.010 | 0.813|$\pm $| 0.014 | 0.390|$\pm $| 0.031 |
This PDF is available to Subscribers Only
View Article Abstract & Purchase OptionsFor full access to this pdf, sign in to an existing account, or purchase an annual subscription.
