The same atom typer for receptor and ligand

[Jalil-Mahdizadeh]

Hi,
I'm trying to generate simple grids for receltor's binding site and it's ligand.
I realized that while I've set "defaultGninaReceptorTyper", the receptor's grid has been generated according to the "defaultGninaLigandTyper".
I'm using molgrid v0.5.1 installed by pip.

import molgrid
from molgrid.openbabel import pybel as pybel
import torch

with open('lig.sdf', 'r') as f:
    lig_content = f.read()

lig = pybel.readstring('sdf', lig_content)
lig_coords = molgrid.CoordinateSet(lig)
lig_center = lig_coords.center() # This is the ligand center

lig_gmaker = molgrid.GridMaker()

lig_dims = lig_gmaker.grid_dimensions(molgrid.defaultGninaLigandTyper.num_types())
lig_gridtensor = torch.zeros(lig_dims, dtype=torch.float32)

lig_gmaker.forward(lig_center, lig_coords, lig_gridtensor)
print("Generated ligand grid shape:", lig_gridtensor.shape)

######################################################################################################

with open('rec.pdb', 'r') as f:
    rec_content = f.read()

rec = pybel.readstring('pdb', rec_content)
rec_coords = molgrid.CoordinateSet(rec)
rec_center = rec_coords.center() # This is the receptor center

recp_gmaker = molgrid.GridMaker()

rec_dims = recp_gmaker.grid_dimensions(molgrid.defaultGninaReceptorTyper.num_types())
rec_gridtensor = torch.zeros(rec_dims, dtype=torch.float32)

recp_gmaker.forward(lig_center, rec_coords, rec_gridtensor)
print("Generated receptor grid shape:", rec_gridtensor.shape)

[dkoes]

When you create the CoordinateSet of the receptor, specify that you want receptor typing:

molgrid.CoordinateSet(r,molgrid.defaultGninaReceptorTyper)

[Jalil-Mahdizadeh]

Great! many thanks.