munkres-rmsd 0.0.1.post4.dev0

Proper RMSD calculation between molecules using the Kuhn-Munkres Hungarian algorithm.

RMSD

Determining the Root Mean Square Deviation (RMSD) between the Maximum Common Substructures (MCS) of two molecules.

The Kuhn-Munkres Hungarian algorithm allows for a fast match of atoms based on:

• atomic symbol
• Sybyl atom types
• pharmacophoric types (i.e. H-bond donors and acceptors or charges)

Installation

pip install munkres-rmsd

Example

from munkres_rmsd import CalcLigRMSD, AtomType
from munkres_rmsd.RMSD import get_example_molecules

# First, load 3D poses of molecules 
mol1, mol2 = get_example_molecules()

# Then compute the RMSD of the best atomic match
rmsd = CalcLigRMSD(mol1, mol2)

print(rmsd) # 10.76150...

Let's use Sybyl atom types to match atoms between the two molecules instead of the default, using atomic elements.

# Then compute the RMSD of the best atomic match
rmsd = CalcLigRMSD(mol1, mol2, AtomType.Sybyl)

print(rmsd) # 11.59752...

Should you prefer pharmacophore types (i.e. H-bond donors & acceptors, charges and others):

# Then compute the RMSD of the best atomic match
rmsd = CalcLigRMSD(mol1, mol2, AtomType.Pharmacophore)

print(rmsd) # 9.49120...

Finally to allow matching any atom whatever its type:

# Then compute the RMSD of the best atomic match
rmsd = CalcLigRMSD(mol1, mol2, AtomType.Skeleton)

print(rmsd) # 8.96288...