- Jupyter Notebook 91.4%
- Python 8.6%
| Filename | Latest commit message | Latest commit date |
|---|---|---|
| .gitignore | ||
| Cyclic peptides disconnection sequences generated.ipynb | ||
| dataset.csv | ||
| experiment.csv | ||
| FeatureExtraction.py | ||
| LICENSE | ||
| Model Selection.ipynb | ||
| Model training.ipynb | ||
| model307.pkl | ||
| Prediction.ipynb | ||
| processed_sequences.csv | ||
| Random Cyclopeptide Generator.ipynb | ||
| README.md | ||
| scaler307.pkl | ||
| Similarity.ipynb | ||
| targets.csv | ||
Machine Learning-Guided Selection of Cyclization Sites in Head-to-Tail Cyclic Peptides
code for training and applying ML model to predict cyclization sites in head-to-tail cyclic peptides
Files
Model Selection.ipynb : iPython notebook for model selection
Model training.ipynb : iPython notebook for training
Prediction.ipynb: iPython notebook for predicting unknown sequences
model307.pkl: best trained model for predicting
scaler307.pkl: unified normalization method
Cyclic peptides disconnection sequences generated.ipynb: iPython notebook for generation of sequences by disconnecting each cyclization site of the cyclic peptides
Random Cyclopeptide Generator: iPython notebook for randomly generation cyclic peptide sequences for experimental validation
Similarity.ipynb: iPython notebook for calcualting sequence similarity
CSV Files
dataset.csv : 306 peptide dataset for training
experiment.csv: 74 sequences for experimental validation
targets.csv: dataset of cyclic peptide sequences designed based on the antibody complementarity-determining regions corresponding to 7 disease targets
processed_sequences.csv: linear peptide sequences generated by disconnecting all cyclization sites of the cyclic peptides designed for the 7 disease targets using the Cyclic peptides disconnection sequences generated .
Install the required packages
conda create -n cyclopeptide python==3.7.1
conda activate cyclopeptide
conda install numpy==1.21.5
conda install pandas==1.3.5
conda install scikit-learn==1.0.2
conda install tensorflow==2.8.2
conda install xgboost==1.5.1
conda install ipykernel
python -m ipykernel install --name cyclopeptide
You can also use pip install, if conda install fails, try using pip install). If using pip install, make sure to install joblib==1.1.1. If a different version is installed, it needs to be uninstalled and reinstalled.
Usage
To run the ML model, navigate to the Anaconda interface and execute the provided Jupyter Notebook (Model training.ipynb). Follow the step-by-step instructions within the notebook to perform the analysis.(Switch the kernel to cyclopeptide in the installation guide)