carbonsilicon-ai / CarsiDock

Official repo of "CarsiDock: a deep learning paradigm for accurate protein–ligand docking and screening based on large-scale pre-training" proposed by CarbonSilicon AI.
http://dx.doi.org/10.1039/D3SC05552C
Apache License 2.0
70 stars 7 forks source link

Docker image reporting errors #3

Closed blurry-cat closed 8 months ago

blurry-cat commented 9 months ago

I've tried many times to install it with the same error. So I would like to ask what is the reason and how to fix it? I think it's probably the error in this step: make install-python-package

#######

DockerFile:28

27 | COPY ./dockcpp /app/dockcpp 28 | >>> RUN cd /app/dockcpp && mkdir build && cd build && \ 29 | >>> cmake .. && make install-python-package && \ 30 | >>> cd ../../ 31 |

ERROR: failed to solve: process "/bin/sh -c cd /app/dockcpp && mkdir build && cd build && cmake .. && make install-python-package && cd ../../" did not complete successfully: exit code: 2

Le-Phung-Hien commented 9 months ago

Hi, I also had the same error. Using the previous version of the dockerfile solved the problem.

blurry-cat commented 9 months ago

Hi, I also had the same error. Using the previous version of the dockerfile solved the problem.

Hi Le, Thank you very much for your attention and reply! I'm not familiar with Github, where do you find the previous version of the dockerfile? Or can you just paste it and provide it, thanks!

gitabtion commented 9 months ago

@blurry-cat Hi, here is the previous vesion of dockerfile:

# https://docs.nvidia.com/deeplearning/frameworks/pytorch-release-notes/rel_22-04.html
FROM nvcr.io/nvidia/pytorch:22.04-py3

RUN conda install -y -c openbabel openbabel
RUN conda install -y -c dglteam dgl-cuda11.6
RUN pip config set global.index-url https://pypi.tuna.tsinghua.edu.cn/simple
COPY ./requirements.txt /app/requirements.txt
RUN pip install -r /app/requirements.txt

COPY ./lib/pydock-0.4-cp38-cp38-manylinux_2_31_x86_64.whl /app/pydock-0.4-cp38-cp38-manylinux_2_31_x86_64.whl
RUN pip install /app/pydock-0.4-cp38-cp38-manylinux_2_31_x86_64.whl

or you can use image in docker hub by changing carsidock:v1 to abtion/carsidock:v2 in command: docker run -v ./:/Docking --gpus all abtion/carsidock:v2 python /Docking/run_core_set_eval.py --cuda_convert

blurry-cat commented 9 months ago

@gitabtion Hi, thanks for your help, now the software is installed!

But I still have some questions about how to use it:

  1. I don't quite understand whether the ①# redocking and ②# docking for decoys in the #Docking step are related to the subsequent #Screening step;
  2. If I download a protein molecule (from the PDB), which already contains its ligand, can I do the Screening step directly? Do I need to remove the ligand structure from the protein before Screening?

Waiting for your reply.

gitabtion commented 9 months ago

@gitabtion Hi, thanks for your help, now the software is installed!

But I still have some questions about how to use it:

  1. I don't quite understand whether the ①# redocking and ②# docking for decoys in the #Docking step are related to the subsequent #Screening step;
  2. If I download a protein molecule (from the PDB), which already contains its ligand, can I do the Screening step directly? Do I need to remove the ligand structure from the protein before Screening?

Waiting for your reply.

@blurry-cat Hi, thanks for your attention.

  1. CarsiDock contains 2 modules: docking and scoring.
    • The docking step (redocking and docking for decoys) only outputs conformers and docking socre (for sorting the conformers). The docking script run_docking_inference.py only contains docking module.
    • The scoring step use rtms to scoring the conformers for screening. The screening script run_screening.py contains the whole process of docking and scoring. The docking results in the above two scripts are the same.
  2. We usually separate the protein and ligand before using CarsiDock.
blurry-cat commented 9 months ago

@gitabtion Hi,thanks again!

Your response addresses the issues I raised. But there are a few more I would like to ask: ①How do I choose the input for the parameter --reflig for protein molecules without a ligand reference (i.e. protein-ligand co-crystallization on the PDB)? ② I tested the Docking step and it outputs 25 conformers by default, is it possible to output only the first conformer after sorting?

Sincerely

gitabtion commented 9 months ago

@gitabtion Hi,thanks again!

Your response addresses the issues I raised. But there are a few more I would like to ask: ①How do I choose the input for the parameter --reflig for protein molecules without a ligand reference (i.e. protein-ligand co-crystallization on the PDB)? ② I tested the Docking step and it outputs 25 conformers by default, is it possible to output only the first conformer after sorting?

Sincerely

@blurry-cat Hi,

  1. CarsiDock is a semi-flexible docking algorithm that requires reflig to cut pockets, and works best when using ligand in the complex crystal as reflig.
  2. Output only the first conformation is not supported yet, you can add a post script for further processing