Table 2.
Currently available TCR-epitope binding prediction methods
| TCR-epitope binding prediction tool | Predictable TCR chain(s) | Epitope constraint | TCR length constraint | Method description | Published date | Software / Webserver |
|---|---|---|---|---|---|---|
| TCRex | TCR β | Restricted epitopes | None | Models based on random forest classifiers | Nov-2019 | Webserver: |
| https://www.tcrex.biodatamining.be/ | ||||||
| ERGO-LSTM | TCR β | None | None | LSTM based model | Aug-2020 | Software: |
| https://github.com/louzounlab/ERGO | ||||||
| ERGO-AE | TCR β | None | None | Autoencoder based model | Aug-2020 | Software: |
| https://github.com/louzounlab/ERGO | ||||||
| ImRex | TCR β | 8 ~ 11-mer | 10 ~ 20 | Dual input CNN model | Dec-2020 | Software: |
| https://github.com/pmoris/ImRex | ||||||
| DLpTCR | TCR αβ | 9-mer | 8 ~ 20 | Ensemble deep learning model of FCN, LeNet-5 and ResNet | Jul-2021 | Software: |
| https://github.com/jiangBiolab/DLpTCR | ||||||
| Webserver: | ||||||
| http://jianglab.org.cn/DLpTCR/ | ||||||
| NetTCR-2.0 | TCR αβ | 9-mer | 8 ~ 18 | 1-dimensional CNN model | Sep-2021 | Software: |
| https://github.com/mnielLab/NetTCR-2.0/ | ||||||
| Webserver: | ||||||
| https://services.healthtech.dtu.dk/service.php?NetTCR-2.0 | ||||||
| pMTnet | TCR β | None | None | deep neural network based on LSTM and stacked autoencoders | Sep-2021 | Software: |
| https://github.com/tianshilu/pMTnet |
Abbreviations LSTM Long short-term memory, CNN,Convolutional neural network, FCN Fully convolutional networks, ResNet Residual neural network