03 Molecular Similarity

Finding the most similar structures to a specific target molecule

import pandas

mols = pandas.read_table('../nci1000.smiles', names=['SMILES', 'NCI_ID'])
mols

	SMILES	NCI_ID
0	CC1=CC(=O)C=CC1=O	1
1	S(SC1=NC2=CC=CC=C2S1)C3=NC4=C(S3)C=CC=C4	2
2	OC1=C(Cl)C=C(C=C1[N+]([O-])=O)[N+]([O-])=O	3
3	[O-][N+](=O)C1=CNC(=N)S1	4
4	NC1=CC2=C(C=C1)C(=O)C3=C(C=CC=C3)C2=O	5
...	...	...
995	OC(=O)CN(CC(O)=O)CC1=CC=CC=C1	1003
996	CC(C)N(CC(O)=O)CC(O)=O	1004
997	C1=CC=C(C=C1)C=NC(N=CC2=CC=CC=C2)C3=CC=CC=C3	1005
998	CC(C)(C)C[C]1(C)NC(=O)NC1=O	1006
999	CCOC(=O)C(C(C)C)C(=O)OCC	1007

1000 rows × 2 columns

from chemaxon.io import import_mol
from chemaxon.fingerprints import ecfp, tanimoto

query = ecfp(import_mol('CC(=O)NC1=CC=C(O)C=C1'), 4, 1024) # ECFP of paracetamol

mols['SIMILARITY'] = mols.apply(lambda row: tanimoto(query, ecfp(import_mol(row['SMILES']), 4, 1024)), axis = 'columns')
mols

	SMILES	NCI_ID	SIMILARITY
0	CC1=CC(=O)C=CC1=O	1	0.117647
1	S(SC1=NC2=CC=CC=C2S1)C3=NC4=C(S3)C=CC=C4	2	0.078947
2	OC1=C(Cl)C=C(C=C1[N+]([O-])=O)[N+]([O-])=O	3	0.121951
3	[O-][N+](=O)C1=CNC(=N)S1	4	0.102564
4	NC1=CC2=C(C=C1)C(=O)C3=C(C=CC=C3)C2=O	5	0.105263
...	...	...	...
995	OC(=O)CN(CC(O)=O)CC1=CC=CC=C1	1003	0.184211
996	CC(C)N(CC(O)=O)CC(O)=O	1004	0.147059
997	C1=CC=C(C=C1)C=NC(N=CC2=CC=CC=C2)C3=CC=CC=C3	1005	0.076923
998	CC(C)(C)C[C]1(C)NC(=O)NC1=O	1006	0.128205
999	CCOC(=O)C(C(C)C)C(=O)OCC	1007	0.111111

1000 rows × 3 columns

mols.hist('SIMILARITY', bins=50)

array([[<Axes: title={'center': 'SIMILARITY'}>]], dtype=object)

most_similars = mols.query('SIMILARITY > 0.5').sort_values(by=['SIMILARITY'], ascending=False)
most_similars

	SMILES	NCI_ID	SIMILARITY
442	CC(=O)NC1=CC=C(Cl)C=C1	444	0.615385
216	CC(=O)NC1=CC=C(C=C1)S(N)(=O)=O	217	0.551724
652	CC(=O)NC1=CC=C(C=C1)C=NO	657	0.548387
634	CC(=O)NC1=CC=C(C=C1)S(=O)(=O)NN	639	0.533333
760	CC(=O)NC1=CC=C(C=C1)C(=O)CCl	768	0.516129

The most similar structure to paracetamol from the target molecules

import_mol(most_similars.iloc[0]['SMILES'])

The second most similar structure

import_mol(most_similars.iloc[1]['SMILES'])