Практикум 2. Хемоинформатика

Загрузим необходимые пакеты

import rdkit
from rdkit import Chem
from rdkit.Chem import AllChem
from rdkit import RDConfig
from rdkit.Chem.Draw import ShowMol
from rdkit.Chem.Draw import IPythonConsole 
from rdkit.Chem import Draw
from IPython.display import display,Image,HTML,Video
from base64 import b64encode
import io
import base64
import numpy as np
import rdkit.Chem.Lipinski as Lipinksy
import pubchempy as pcp

Молекула ибупрофена:

ibu=Chem.MolFromSmiles('CC(C)CC1=CC=C(C=C1)C(C)C(=O)O')
AllChem.Compute2DCoords(ibu)
display(ibu)

print(Lipinksy.NumHDonors(ibu))
print(Lipinksy.NumHAcceptors(ibu))
print(Lipinksy.rdMolDescriptors.CalcExactMolWt(ibu))
print(Lipinksy.rdMolDescriptors.CalcCrippenDescriptors(ibu)[0])
def is_valid_lipinsky(molecule):
    return bool((Lipinksy.NumHDonors(molecule) <= 5)
                and (Lipinksy.NumHAcceptors(molecule) <= 10)
                and (Lipinksy.rdMolDescriptors.CalcExactMolWt(molecule) < 500)
                and(Lipinksy.rdMolDescriptors.CalcCrippenDescriptors(molecule)[0]) <= 5)
print(is_valid_lipinsky(ibu))

1
1
206.130679816
3.073200000000001
True

Ибупрофен, модифицированный для Click Chemistry:

ibu1 = Chem.MolFromSmiles('CC(C#C)CC1=CC=C(C=C1)C(C)C(=O)O')
tr1 = Chem.MolFromSmiles('N1N=NC(C(C)CC2=CC=C(C=C2)C(C)C(=O)O)=C1')

AllChem.Compute2DCoords(ibu1)
AllChem.Compute2DCoords(tr1)
display(ibu1,tr1)

compounds = []
per_page = 10**5
for smiles in ["N=N=N", "NN#N"]:
    for i in range(200):
        try:
            a = pcp.get_properties(
              properties="CanonicalSMILES", 
              identifier=smiles, namespace='smiles', 
              searchtype="substructure",
              RingsNotEmbedded=True,
              listkey_count=per_page, listkey_start=i*per_page
            )
        except:
            continue
        print("Retrieved page {} of {} search".format(i+1, smiles))
        compounds.extend(a)

Retrieved page 1 of N=N=N search
Retrieved page 2 of N=N=N search
Retrieved page 3 of N=N=N search
Retrieved page 4 of N=N=N search
Retrieved page 5 of N=N=N search
Retrieved page 6 of N=N=N search
Retrieved page 7 of N=N=N search
Retrieved page 8 of N=N=N search
Retrieved page 9 of N=N=N search
Retrieved page 10 of N=N=N search
Retrieved page 11 of N=N=N search
Retrieved page 12 of N=N=N search
Retrieved page 13 of N=N=N search
Retrieved page 14 of N=N=N search
Retrieved page 15 of N=N=N search
Retrieved page 16 of N=N=N search
Retrieved page 19 of N=N=N search
Retrieved page 1 of NN#N search

mols = []

for smi in compounds:
    if 'N=[N+]=[N-]' in list(smi.values())[1]:
        newsmi = list(smi.values())[1].replace('N=[N+]=[N-]', 'CC(C)CC1=CC=C(C=C1)C(C)C(=O)O')
    else:
        continue

    try:
        newmol = Chem.MolFromSmiles(newsmi)
        if is_valid_lipinsky(newmol):          
            mols.append(newmol)
    except:
        pass


len(mols)

551935

display(Draw.MolsToGridImage(
    mols,
    maxMols=9,
    molsPerRow=3,
    subImgSize=(400, 400)))

Similiraty Map ибупрофена с веществом из массива:

SimilarityMaps.GetSimilarityMapForFingerprint(
        ibu,
        mols[0],
        SimilarityMaps.GetMorganFingerprint
    )

(<Figure size 180x180 with 1 Axes>, 0.18729096989966554)

m3d = Chem.AddHs(mols[0])
Chem.AllChem.EmbedMolecule(m3d)
AllChem.MMFFOptimizeMolecule(m3d, maxIters=500, nonBondedThresh=200)
display(m3d)

import nglview as nv
nv.show_rdkit(m3d)