LigParGen server 2.0, Updated LigParGen version

To install LigParGen follow these steps:

• Get the latest version of BOSS
  • Download it and uncompress it into this directory
• Either create a new environment conda create --name ligpargen or activate your HiTPoly (or other) environment
• Install dependencies
  • conda install openbabel -c openbabel
• Clone this repo onto your directory
• Navigate in the directory and pip install -e .
• Add to your .bashrc/.zshrc the following line
  • export BOSSdir=$HOME/ligpargen/boss
• You can also export a ligpargen path and call it as
  • export LigParGen=$HOME/ligpargen/LigParGen
  • python $LigParGen/Converter.py and then the ligpargen command
• LigParGen uses a c-like syntax interpreter, so you'll have to ask the superuser on a machine to install csh if it's already not
  • sudo apt-get install csh
• This worked with python up to 3.10 (and potentially more)

The main script is compiled on a 32-bit version so there might be sometimes problems for some machines that don't have Multi-Arhc support enabled by default, if that's the case, you must resort to using containers!

Okay, everything now runs on podman/docker

• You need to have podman/docker on your workstation installed to have this work, podman, docker
• Important!! For images to work you must uncomment the os.environ["BOSSdir"]... line in the LigParGen/Converter.py file
• Run the following command, this will build the container image
  • podman build -t ligpargen .
• docker build -t ligpargen .
• If you are on an apple silicon machine or ARM-based linux one you need to ensure the image works for Intel-based environments, so run either one of these:
  • podman build --platform linux/amd64 -t ligpargen .
  • podman build --platform linux/amd64 -t ligpargen .
• Then you must save the container image in some directory via
  • podman save -o ./ligpargen.tar ligpargen:latest
  • docker save -o ./ligpargen.tar ligpargen:latest
• Then you must move the ligpargen.tar file to your cluster in a folder in your home directory called containers, and then everything else is done inside hitpoly (with platform set to 'supercloud').

To run LigParGen

This is a test case for a simple glyme molecule. Create and run a python script with this code with your desired smiles in the ./RUN directory and then run the commands below. The ./RUN directory already has the necessary file to test if LigParGen works.

from rdkit import Chem
from rdkit.Chem import AllChem

smiles = "COCCOC"
mol = Chem.MolFromSmiles(smiles)
mol = Chem.AddHs(mol, addCoords=True)

initial_confs = AllChem.EmbedMultipleConfs(
mol,
numConfs=1,
maxAttempts=50000,
boxSizeMult=5,
useRandomCoords=False, # use 'True' here if this doesn't generate a 3D conformation, LigParGen will fail if it doesn't have a 3D conformation
)

print(Chem.MolToMolBlock(mol),file=open(f'./poly.mol','w+'))

And then run the command in the same folder where you've saved the poly.mol file

python ../LigParGen/Converter.py -m poly.mol -o 0 -c 0 -r PLY -d . -l

or

$LigParGen/Converter.py -m poly.mol -o 0 -c 0 -r PLY -d . -l

Official page/post about the local version of ligpargen

Official post about the dockerized version of ligpargen

Info from the original github repo