pengsihua2023 / bindppi Goto Github PK

BindPPI calculates binding affinities for protein-protein interactions. It includes three representative best-performing models. The first model is structure-based, consisting of a random forest regression and thirteen handcrafted features, named as RF_{13}. The second model is sequence-based, employing an architecture that combines transferred embedding features with a multilayer perceptron, named as MLP_{5120}. Finally, we created an ensemble model by averaging the predictions of the two aforementioned models, named as AvgEns. For RF_{13} and AvgEns, the input requires the 3D structure of the protein-protein complex, while MLP_{5120} only needs the sequence of the complex for computation.

The prefered and easiest is by pulling the docker image made available publicly.

prerequisites: docker

Just pull the image:

docker pull minghuilab/bindppi:v1

Note: this step may take about 20mins.

docker run -it minghuilab/bindppi:v1 /bin/bash
conda activate myenv
cd bindppi

python BindPPI_RF13.py -i example_1akj

Note: the folder name (example_1akj) and input file name (example_1akj.input) should be in lowercase, and their file names should match.

python BindPPI_MLP5120.py -f ./MLP5120_example/sample_input_sequence.json -o your_output_file

To run the MLP_{5120} model, you only need to create a JSON file containing the input sequences for each complex (refer to the file sample_input_sequence.json in the MLP5120_example folder). The output file is your_output_file, referring to the example MLP5120_prediction.txt.

python BindPPI_AvgEns.py -i example_1akj

To run the AvgEns model, you need to (1) create a folder to organize your input and output files. Let's call it example_1akj; (2) prepare three input files similar to example_1akj.input, 1AKJ.pdb, and example_1akj.json in the example_1akj folder. The output file, example_1akj.AvgEns, will be generated as a result of running the model.