def __init__(self, reference: str = None, target_s: str = None):
'''
Setup of the SuCOS scoring class. If you do not set reference or target_s, you will have to feed them
as rdkit mol objects to each class method/function.
Sets: self.reference: an rdkit mol object of the reference mol
self.target_s: a list of rdkit mol objects for the target mols
:param reference: an sdf or mol file containing the reference (hit) molecule
:param target_s: an sdf or mol file containing one or multiple molecules to compare against reference
'''
self.fdef = AllChem.BuildFeatureFactory(
os.path.join(RDConfig.RDDataDir, 'BaseFeatures.fdef'))
self.fmParams = {}
for k in self.fdef.GetFeatureFamilies():
self.fparams = FeatMaps.FeatMapParams()
self.fmParams[k] = self.fparams
self.keep = ('Donor', 'Acceptor', 'NegIonizable', 'PosIonizable',
'ZnBinder', 'Aromatic', 'Hydrophobe', 'LumpedHydrophobe')
if reference:
refs = Chem.SDMolSupplier(reference, sanitize=True)
self.reference = [x for x in refs if x][0]
if target_s:
tgts = Chem.SDMolSupplier(target_s, sanitize=True)
self.target_s = [x for x in tgts if x]
def donor_acceptor(mol):
fdef = AllChem.BuildFeatureFactory(
os.path.join(RDConfig.RDDataDir, 'BaseFeatures.fdef'))
mol_feat = fdef.GetFeaturesForMol(mol)
idx = [atom_feat.GetAtomIds()[0] for atom_feat in mol_feat]
mol_idx = [atom.GetIdx() for atom in mol.GetAtoms()]
k = list(range(mol.GetNumAtoms()))
a = [i for i in mol_idx if i not in idx]
for i, atom_feat in zip(idx, mol_feat):
k[i] = atom_feat.GetFamily()
for i in a:
k[i] = "None"
return k
def show_pharmacophore(sdf_path,
keep=keep,
fdf_dir=os.path.join(os.path.dirname(__file__),
'defined_BaseFeatures.fdef')):
template_mol = [m for m in Chem.SDMolSupplier(sdf_path)][0]
fdef = AllChem.BuildFeatureFactory(fdf_dir)
prob_feats = fdef.GetFeaturesForMol(template_mol)
prob_feats = [f for f in prob_feats if f.GetFamily() in keep]
# prob_points = [list(x.GetPos()) for x in prob_feats]
for i, feat in enumerate(prob_feats):
atomids = feat.GetAtomIds()
print("pharamcophore index:{0}; feature:{1}; type:{2}; atom id:{3}".
format(i, feat.GetFamily(), feat.GetType(), atomids))
display(
Draw.MolToImage(template_mol,
highlightAtoms=list(atomids),
highlightColor=[0, 1, 0],
useSVG=True))
Publication: https://doi.org/10.26434/chemrxiv.8100203.v1
"""
from __future__ import print_function
import argparse, os, sys, gzip
import numpy as np
from rdkit import Chem, rdBase, RDConfig
from rdkit.Chem import AllChem, rdShapeHelpers
from rdkit.Chem.FeatMaps import FeatMaps
import utils
### start function definitions #########################################
# Setting up the features to use in FeatureMap
fdef = AllChem.BuildFeatureFactory(os.path.join(RDConfig.RDDataDir, 'BaseFeatures.fdef'))
fmParams = {}
for k in fdef.GetFeatureFamilies():
fparams = FeatMaps.FeatMapParams()
fmParams[k] = fparams
keep = ('Donor', 'Acceptor', 'NegIonizable', 'PosIonizable', 'ZnBinder',
'Aromatic', 'Hydrophobe', 'LumpedHydrophobe')
def filterFeature(f):
result = f.GetFamily() in keep
# TODO - nothing ever seems to be filtered. Is this expected?
if not result:
utils.log("Filtered out feature type", f.GetFamily())
return result
import os
from rdkit import Chem
from rdkit.Chem import AllChem, rdShapeHelpers
from rdkit.Chem.FeatMaps import FeatMaps
from rdkit import RDConfig
# Set up features to use in FeatureMap
fdefName = os.path.join(RDConfig.RDDataDir, 'BaseFeatures.fdef')
fdef = AllChem.BuildFeatureFactory(fdefName)
fmParams = {}
for k in fdef.GetFeatureFamilies():
fparams = FeatMaps.FeatMapParams()
fmParams[k] = fparams
keep = ('Donor', 'Acceptor', 'NegIonizable', 'PosIonizable', 'ZnBinder',
'Aromatic', 'Hydrophobe', 'LumpedHydrophobe')
def get_FeatureMapScore(query_mol, ref_mol):
featLists = []
for m in [query_mol, ref_mol]:
rawFeats = fdef.GetFeaturesForMol(m)
# filter that list down to only include the ones we're intereted in
featLists.append([f for f in rawFeats if f.GetFamily() in keep])
fms = [
FeatMaps.FeatMap(feats=x, weights=[1] * len(x), params=fmParams)
for x in featLists
]
fms[0].scoreMode = FeatMaps.FeatMapScoreMode.Best
fm_score = fms[0].ScoreFeats(featLists[1]) / min(fms[0].GetNumFeatures(),
