Deep Learning Method for Anticancer Peptide Activity Prediction through Convolutional Neural Network and Multi-Task Learning
🔥UPDATE(2021.7.12): The online server for xDeep-AcPEP is now available: Here
In this repository:
- We provide python scripts for reproducing the experiments of 5-folds cross-validation model comparison.
- We provide our final production models for peptide activity prediction.
- Anaconda 4.7.0
- Python 3.6.9
- Scikit-learn 0.21.3
- Pytorch 1.2.0 with CUDA 10.0
- Scipy 1.4.1
- Pandas 1.0.2
- Numpy 1.18.1
- Openpyxl 3.0.6
- Models and data used for reproducing experiments are available at: Here
- Final production models for peptide activity prediction are available at:
./prediction/model/
The script is located in model_comparison_CV
folder
python reproduce.py -mo <model_folder_path> -da <data_folder_path> -o <output_folder_path>
python reproduce.py -mo ./model/ -da ./data/ -o ./
2. Final production model prediction (Online Server)
The script is located in prediction
folder
python prediction.py -t <tissue_type> -m <model_folder_path> -d <fasta_file_path> -o <output_folder_path>
where:
<tissue_type>
could be selected from breast
, cervix
, colon
, lung
, prostate
and skin
.
python prediction.py -t breast -m ./model/ -d ./test_breast.fasta -o ./result/
Note: The input peptide data must in the form of the following FASTA format.
>AmphiArc1
KWVKKVHNWLRRWIKVFEALFG
>AmphiArc2
KIFKKFKTIIKKVWRIFGRF
>Gradient2
AWLKRIKKFLKALFWVWVW
>AmphiArc3
AFRHSVKEELNYIRRRLERFPNRL