externelly/chainsaw repository overview

⁠Chainsaw Docker Image

This Docker image is meant to run Chainsaw⁠, a package for protein domain segmentation based on a given PDB structure. Create a mount/ and add single PDB files or a folder containing PDB files to into so the results will properly save once the analysis is complete. Takes only a few seconds to run each PDB, with the GPU providing ~5-10x speedup. Remove the --gpus all if you don't want to use GPUs for the analysis. You can find the Dockerfile on my GitHub repository⁠ and additional information on my website Docker page⁠.

docker pull externelly/chainsaw:latest

# To run the Chainsaw example file
docker run --gpus all -it -v ${PWD}/mount:/chainsaw/results/ externelly/chainsaw:latest --structure_file example_files/AF-A0A1W2PQ64-F1-model_v4.pdb --output results/AF-A0A1W2PQ64-F1-model_v4.tsv

# To run Chainsaw on a folder containing multiple PDB files
docker run --gpus all -it -v ${PWD}/mount:/chainsaw/mount/ externelly/chainsaw:latest --structure_directory mount/pdbs --output mount/pdbs.tsv

# To run Chainsaw on a single PDB file
docker run --gpus all -it -v ${PWD}/mount:/chainsaw/mount/ externelly/chainsaw:latest --structure_file mount/AF-A0A1D6DZU6-F1-model_v4.pdb --output mount/AF-A0A1D6DZU6-F1-model_v4.tsv

⁠Cite

Don't forget to cite the original paper:

Wells, J., Hawkins-Hooker, A., Bordin, N., Paige, B., & Orengo, C. (2023). Chainsaw: protein domain segmentation with fully convolutional neural networks. bioRxiv.