Welcome to ATPfinder, a deep forest-based ATP prediction tool developed by a team from the Chinese University of Hong Kong (Shenzhen)
ATPfinder is a computational framework for identifying anti-tubercular peptides via deep forest architecture with effective feature representation. We also developed a downloadable desktop program for ATPfinder, which is available at https://awi.cuhk.edu.cn/~dbAMP/ATPfinder.html.
First you need to install the Python environment for Deep Forest, via the following command.
pip install deep-forest
For details of this python package, please refer to the API here.
- The dataset used for the model can be found here: https://github.com/lantianyao/ATPfinder/tree/main/data
- Feature extraction.ipynb provides code to extract peptide descriptors used in this study.
- Model Training.ipynb is a quick tutorial that tells you how to train deep forest-based model with cross-validation.
- Using the trained model.ipynb is a quick tutorial telling you how to use a trained deep forest model.
If you have any questions, please feel free to contact me.
Name: Lantian Yao Email: [email protected]