{
"cells": [
{
"cell_type": "code",
"execution_count": 1,
"id": "95bd761a-fe51-4a8e-bc70-1365260ba5f8",
"metadata": {},
"outputs": [],
"source": [
"import pandas as pd"
]
},
{
"cell_type": "code",
"execution_count": 2,
"id": "b0859483-5e19-4280-9f53-0d00a6f22d34",
"metadata": {},
"outputs": [],
"source": [
"df_pdbbind = pd.read_parquet('data/pdbbind.parquet')\n",
"df_pdbbind = df_pdbbind[['seq','smiles','affinity_uM']]"
]
},
{
"cell_type": "code",
"execution_count": 3,
"id": "f30732b7-7444-47ad-84e7-566e7a6f2f8e",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"
\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" seq | \n",
" smiles | \n",
" affinity_uM | \n",
"
\n",
" \n",
" \n",
" \n",
" 0 | \n",
" MTVPDRSEIAGKWYVVALASNTEFFLREKDKMKMAMARISFLGEDE... | \n",
" CCCCCCCCCCCCCCCCCCC[C-](=O)=O | \n",
" 0.026 | \n",
"
\n",
" \n",
" 1 | \n",
" AAPFDKSKNVAQSIDQLIGQTPALYLNKLNNTKAKVVLKMECENPM... | \n",
" OC[C@@H](C(=O)N[C@@H]([C@H](CC)C)[C-](=O)=O)NC... | \n",
" 6.430 | \n",
"
\n",
" \n",
" 2 | \n",
" PFPLTSMDKAFITVLEMTPVLGTEIINYRDGMGRVLAQDVYAKDNL... | \n",
" CC[C@@H]([C@@H](C(=O)N[C@H](C(=O)NCC(=O)N[C@H]... | \n",
" 190.000 | \n",
"
\n",
" \n",
" 3 | \n",
" SMENFQKVEKIGEGTYGVVYKARNKLTGEVVALKKIRLDTETEGVP... | \n",
" OCC[C@@H]1CCCCN1c1cc(NCC2=CC=CN(C2)O)n2c(n1)c(... | \n",
" 0.210 | \n",
"
\n",
" \n",
" 4 | \n",
" EFSEWFHNILEEAEIIDQRYPVKGMHVWMPHGFMIRKNTLKILRRI... | \n",
" O[C@@H]1[C@@H](COS(=O)(=O)NC(=O)[C@@H]2CCC[NH2... | \n",
" 0.050 | \n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" seq \\\n",
"0 MTVPDRSEIAGKWYVVALASNTEFFLREKDKMKMAMARISFLGEDE... \n",
"1 AAPFDKSKNVAQSIDQLIGQTPALYLNKLNNTKAKVVLKMECENPM... \n",
"2 PFPLTSMDKAFITVLEMTPVLGTEIINYRDGMGRVLAQDVYAKDNL... \n",
"3 SMENFQKVEKIGEGTYGVVYKARNKLTGEVVALKKIRLDTETEGVP... \n",
"4 EFSEWFHNILEEAEIIDQRYPVKGMHVWMPHGFMIRKNTLKILRRI... \n",
"\n",
" smiles affinity_uM \n",
"0 CCCCCCCCCCCCCCCCCCC[C-](=O)=O 0.026 \n",
"1 OC[C@@H](C(=O)N[C@@H]([C@H](CC)C)[C-](=O)=O)NC... 6.430 \n",
"2 CC[C@@H]([C@@H](C(=O)N[C@H](C(=O)NCC(=O)N[C@H]... 190.000 \n",
"3 OCC[C@@H]1CCCCN1c1cc(NCC2=CC=CN(C2)O)n2c(n1)c(... 0.210 \n",
"4 O[C@@H]1[C@@H](COS(=O)(=O)NC(=O)[C@@H]2CCC[NH2... 0.050 "
]
},
"execution_count": 3,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df_pdbbind.head()"
]
},
{
"cell_type": "code",
"execution_count": 4,
"id": "2787b9fd-3d6f-4ae3-a3ad-d3539b72782b",
"metadata": {},
"outputs": [],
"source": [
"from rdkit import Chem\n",
"from rdkit.Chem import MACCSkeys\n",
"import numpy as np\n",
"\n",
"def get_maccs(smi):\n",
" try:\n",
" mol = Chem.MolFromSmiles(smi)\n",
" arr = np.packbits([0 if c=='0' else 1 for c in MACCSkeys.GenMACCSKeys(mol).ToBitString()])\n",
" return np.pad(arr,(0,3)).view(np.uint32)\n",
" except Exception:\n",
" pass"
]
},
{
"cell_type": "code",
"execution_count": 6,
"id": "d1abe1c8-ac66-4289-8964-367a5b18528d",
"metadata": {},
"outputs": [],
"source": [
"df_bindingdb = pd.read_parquet('data/bindingdb.parquet')\n",
"df_bindingdb = df_bindingdb[['seq','Ligand SMILES','affinity_uM']].rename(columns={'Ligand SMILES': 'smiles'})"
]
},
{
"cell_type": "code",
"execution_count": 7,
"id": "988bab9c-5147-44e2-92ef-902eaf3c5a90",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" seq | \n",
" smiles | \n",
" affinity_uM | \n",
"
\n",
" \n",
" \n",
" \n",
" 0 | \n",
" PQITLWQRPLVTIKIGGQLKEALLDTGADDTVLEEMSLPGRWKPKM... | \n",
" COc1cc2c(Nc3ccc(Br)cc3F)ncnc2cc1OCC1CCN(C)CC1 | \n",
" 0.00024 | \n",
"
\n",
" \n",
" 1 | \n",
" PQITLWQRPLVTIKIGGQLKEALLDTGADDTVLEEMSLPGRWKPKM... | \n",
" O[C@@H]1[C@@H](O)[C@@H](Cc2ccccc2)N(C\\C=C\\c2cn... | \n",
" 0.00025 | \n",
"
\n",
" \n",
" 2 | \n",
" PQITLWQRPLVTIKIGGQLKEALLDTGADDTVLEEMSLPGRWKPKM... | \n",
" O[C@@H]1[C@@H](O)[C@@H](Cc2ccccc2)N(CC2CC2)C(=... | \n",
" 0.00041 | \n",
"
\n",
" \n",
" 3 | \n",
" PQITLWQRPLVTIKIGGQLKEALLDTGADDTVLEEMSLPGRWKPKM... | \n",
" OCCCCCCN1[C@H](Cc2ccccc2)[C@H](O)[C@@H](O)[C@@... | \n",
" 0.00080 | \n",
"
\n",
" \n",
" 4 | \n",
" PQITLWQRPLVTIKIGGQLKEALLDTGADDTVLEEMSLPGRWKPKM... | \n",
" OCCCCCN1[C@H](Cc2ccccc2)[C@H](O)[C@@H](O)[C@@H... | \n",
" 0.00099 | \n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" seq \\\n",
"0 PQITLWQRPLVTIKIGGQLKEALLDTGADDTVLEEMSLPGRWKPKM... \n",
"1 PQITLWQRPLVTIKIGGQLKEALLDTGADDTVLEEMSLPGRWKPKM... \n",
"2 PQITLWQRPLVTIKIGGQLKEALLDTGADDTVLEEMSLPGRWKPKM... \n",
"3 PQITLWQRPLVTIKIGGQLKEALLDTGADDTVLEEMSLPGRWKPKM... \n",
"4 PQITLWQRPLVTIKIGGQLKEALLDTGADDTVLEEMSLPGRWKPKM... \n",
"\n",
" smiles affinity_uM \n",
"0 COc1cc2c(Nc3ccc(Br)cc3F)ncnc2cc1OCC1CCN(C)CC1 0.00024 \n",
"1 O[C@@H]1[C@@H](O)[C@@H](Cc2ccccc2)N(C\\C=C\\c2cn... 0.00025 \n",
"2 O[C@@H]1[C@@H](O)[C@@H](Cc2ccccc2)N(CC2CC2)C(=... 0.00041 \n",
"3 OCCCCCCN1[C@H](Cc2ccccc2)[C@H](O)[C@@H](O)[C@@... 0.00080 \n",
"4 OCCCCCN1[C@H](Cc2ccccc2)[C@H](O)[C@@H](O)[C@@H... 0.00099 "
]
},
"execution_count": 7,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df_bindingdb.head()"
]
},
{
"cell_type": "code",
"execution_count": 8,
"id": "d7bfee2a-c4e6-48c9-b0c6-52f6a69c7453",
"metadata": {},
"outputs": [],
"source": [
"df_moad = pd.read_parquet('data/moad.parquet')\n",
"df_moad = df_moad[['seq','smiles','affinity_uM']]"
]
},
{
"cell_type": "code",
"execution_count": 9,
"id": "25553199-1715-40fb-9260-427bdd6c3706",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" seq | \n",
" smiles | \n",
" affinity_uM | \n",
"
\n",
" \n",
" \n",
" \n",
" 0 | \n",
" NYIVPGEYRVAEGEIEINAGREKTTIRVSNTGDRPIQVGSHIHFVE... | \n",
" NP(=O)(N)O | \n",
" 0.000620 | \n",
"
\n",
" \n",
" 1 | \n",
" NYIVPGEYRVAEGEIEINAGREKTTIRVSNTGDRPIQVGSHIHFVE... | \n",
" CC(=O)NO | \n",
" 2.600000 | \n",
"
\n",
" \n",
" 2 | \n",
" MTDMSIKFELIDVPIPQGTNVIIGQAHFIKTVEDLYEALVTSVPGV... | \n",
" c1nc(c2c(n1)n(cn2)[C@H]3[C@@H]([C@@H]([C@H](O3... | \n",
" 15.000000 | \n",
"
\n",
" \n",
" 3 | \n",
" MTDMSIKFELIDVPIPQGTNVIIGQAHFIKTVEDLYEALVTSVPGV... | \n",
" c1nc(c2c(n1)n(cn2)[C@H]3[C@@H]([C@@H]([C@H](O3... | \n",
" 15.000000 | \n",
"
\n",
" \n",
" 4 | \n",
" MTDMSIKFELIDVPIPQGTNVIIGQAHFIKTVEDLYEALVTSVPGV... | \n",
" c1nc(c2c(n1)n(cn2)[C@H]3[C@@H]([C@@H]([C@H](O3... | \n",
" 15.000000 | \n",
"
\n",
" \n",
" ... | \n",
" ... | \n",
" ... | \n",
" ... | \n",
"
\n",
" \n",
" 17682 | \n",
" MGSSHHHHHHSSGLVPRGSHMASNPSLIRSESWQVYEGNEANLLDG... | \n",
" None | \n",
" 127.226463 | \n",
"
\n",
" \n",
" 17683 | \n",
" MGSSHHHHHHSSGLVPRGSHMASNPSLIRSESWQVYEGNEANLLDG... | \n",
" None | \n",
" 127.226463 | \n",
"
\n",
" \n",
" 17684 | \n",
" MGSSHHHHHHSSGLVPRGSHMASNPSLIRSESWQVYEGNEANLLDG... | \n",
" None | \n",
" 169.204738 | \n",
"
\n",
" \n",
" 17685 | \n",
" MGSSHHHHHHSSGLVPRGSHMASNPSLIRSESWQVYEGNEANLLDG... | \n",
" None | \n",
" 169.204738 | \n",
"
\n",
" \n",
" 17686 | \n",
" MGSSHHHHHHSSGLVPRGSHMASNPSLIRSESWQVYEGNEANLLDG... | \n",
" None | \n",
" 169.204738 | \n",
"
\n",
" \n",
"
\n",
"
17687 rows × 3 columns
\n",
"
"
],
"text/plain": [
" seq \\\n",
"0 NYIVPGEYRVAEGEIEINAGREKTTIRVSNTGDRPIQVGSHIHFVE... \n",
"1 NYIVPGEYRVAEGEIEINAGREKTTIRVSNTGDRPIQVGSHIHFVE... \n",
"2 MTDMSIKFELIDVPIPQGTNVIIGQAHFIKTVEDLYEALVTSVPGV... \n",
"3 MTDMSIKFELIDVPIPQGTNVIIGQAHFIKTVEDLYEALVTSVPGV... \n",
"4 MTDMSIKFELIDVPIPQGTNVIIGQAHFIKTVEDLYEALVTSVPGV... \n",
"... ... \n",
"17682 MGSSHHHHHHSSGLVPRGSHMASNPSLIRSESWQVYEGNEANLLDG... \n",
"17683 MGSSHHHHHHSSGLVPRGSHMASNPSLIRSESWQVYEGNEANLLDG... \n",
"17684 MGSSHHHHHHSSGLVPRGSHMASNPSLIRSESWQVYEGNEANLLDG... \n",
"17685 MGSSHHHHHHSSGLVPRGSHMASNPSLIRSESWQVYEGNEANLLDG... \n",
"17686 MGSSHHHHHHSSGLVPRGSHMASNPSLIRSESWQVYEGNEANLLDG... \n",
"\n",
" smiles affinity_uM \n",
"0 NP(=O)(N)O 0.000620 \n",
"1 CC(=O)NO 2.600000 \n",
"2 c1nc(c2c(n1)n(cn2)[C@H]3[C@@H]([C@@H]([C@H](O3... 15.000000 \n",
"3 c1nc(c2c(n1)n(cn2)[C@H]3[C@@H]([C@@H]([C@H](O3... 15.000000 \n",
"4 c1nc(c2c(n1)n(cn2)[C@H]3[C@@H]([C@@H]([C@H](O3... 15.000000 \n",
"... ... ... \n",
"17682 None 127.226463 \n",
"17683 None 127.226463 \n",
"17684 None 169.204738 \n",
"17685 None 169.204738 \n",
"17686 None 169.204738 \n",
"\n",
"[17687 rows x 3 columns]"
]
},
"execution_count": 9,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df_moad"
]
},
{
"cell_type": "code",
"execution_count": 10,
"id": "b2c936bc-cdc8-4bc1-b92d-f8755fd65f0a",
"metadata": {},
"outputs": [],
"source": [
"df_biolip = pd.read_parquet('data/biolip.parquet')\n",
"df_biolip = df_biolip[['seq','smiles','affinity_uM']]"
]
},
{
"cell_type": "code",
"execution_count": 11,
"id": "cee93018-601d-458b-af44-bd978da7a2bc",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" seq | \n",
" smiles | \n",
" affinity_uM | \n",
"
\n",
" \n",
" \n",
" \n",
" 38 | \n",
" PYTVVYFPVRGRCAALRMLLADQGQSWKEEVVTVETWQEGSLKASC... | \n",
" CC[C@H](C(=O)c1ccc(c(c1Cl)Cl)OCC(=O)O)C | \n",
" 1.5000 | \n",
"
\n",
" \n",
" 43 | \n",
" MPPYTVVYFPVRGRCAALRMLLADQGQSWKEEVVTVETWQEGSLKA... | \n",
" OC(=O)c1cc(/N=N/c2ccc(cc2)S(=O)(=O)Nc2ccccn2)c... | \n",
" 24.0000 | \n",
"
\n",
" \n",
" 53 | \n",
" EKKSINECDLKGKKVLIRVDFNVPVKNGKITNDYRIRSALPTLKKV... | \n",
" O[C@@H]1[C@@H](CO[P@](=O)(O[P@@](=O)(C(CCCC(P(... | \n",
" 6.0000 | \n",
"
\n",
" \n",
" 54 | \n",
" MPPYTVVYFPVRGRCAALRMLLADQGQSWKEEVVTVETWQEGSLKA... | \n",
" CCCCCCSC[C@@H](C(=O)NCC(=O)O)NC(=O)CC[C@@H](C(... | \n",
" 10.0000 | \n",
"
\n",
" \n",
" 55 | \n",
" MNIFEMLRIDEGLRLKIYKDTEGYYTIGIGHLLTKSPSLNAAKSEL... | \n",
" c1ccccc1 | \n",
" 175.0000 | \n",
"
\n",
" \n",
" ... | \n",
" ... | \n",
" ... | \n",
" ... | \n",
"
\n",
" \n",
" 105118 | \n",
" PQITLWQRPLVTIKIGGQLKEALLDTGADDTVLEEMNLPGRWKPKM... | \n",
" O[C@@H]([C@H](Cc1ccccc1)NC(=O)[C@H](C(C)C)NC(=... | \n",
" 0.0045 | \n",
"
\n",
" \n",
" 105119 | \n",
" PQITLWQRPLVTIKIGGQLKEALLDTGADDTVLEEMNLPGRWKPKM... | \n",
" O[C@@H]([C@H](Cc1ccccc1)NC(=O)[C@H](C(C)C)NC(=... | \n",
" 0.0045 | \n",
"
\n",
" \n",
" 105124 | \n",
" SKVVVPAQGKKITLQNGKLNVPENPIIPYIEGDGIGVDVTPAMLKV... | \n",
" O[C@@H]1[C@@H](COP(=O)(O)O)O[C@H]([C@@H]1OP(=O... | \n",
" 125.0000 | \n",
"
\n",
" \n",
" 105133 | \n",
" ANIVGGIEYSINNASLCSVGFSVTRGATKGFVTAGHCGTVNATARI... | \n",
" CC(C[C@@H](B(O)O)NC(=O)[C@@H]1CCCN1C(=O)[C@@H]... | \n",
" 2.0000 | \n",
"
\n",
" \n",
" 105138 | \n",
" KFPRVKNWELGSITYDTLCAQSQQDGPCTPRRCLGSLVLPRKLQTR... | \n",
" CC[Se]C(=N)N | \n",
" 0.0390 | \n",
"
\n",
" \n",
"
\n",
"
13645 rows × 3 columns
\n",
"
"
],
"text/plain": [
" seq \\\n",
"38 PYTVVYFPVRGRCAALRMLLADQGQSWKEEVVTVETWQEGSLKASC... \n",
"43 MPPYTVVYFPVRGRCAALRMLLADQGQSWKEEVVTVETWQEGSLKA... \n",
"53 EKKSINECDLKGKKVLIRVDFNVPVKNGKITNDYRIRSALPTLKKV... \n",
"54 MPPYTVVYFPVRGRCAALRMLLADQGQSWKEEVVTVETWQEGSLKA... \n",
"55 MNIFEMLRIDEGLRLKIYKDTEGYYTIGIGHLLTKSPSLNAAKSEL... \n",
"... ... \n",
"105118 PQITLWQRPLVTIKIGGQLKEALLDTGADDTVLEEMNLPGRWKPKM... \n",
"105119 PQITLWQRPLVTIKIGGQLKEALLDTGADDTVLEEMNLPGRWKPKM... \n",
"105124 SKVVVPAQGKKITLQNGKLNVPENPIIPYIEGDGIGVDVTPAMLKV... \n",
"105133 ANIVGGIEYSINNASLCSVGFSVTRGATKGFVTAGHCGTVNATARI... \n",
"105138 KFPRVKNWELGSITYDTLCAQSQQDGPCTPRRCLGSLVLPRKLQTR... \n",
"\n",
" smiles affinity_uM \n",
"38 CC[C@H](C(=O)c1ccc(c(c1Cl)Cl)OCC(=O)O)C 1.5000 \n",
"43 OC(=O)c1cc(/N=N/c2ccc(cc2)S(=O)(=O)Nc2ccccn2)c... 24.0000 \n",
"53 O[C@@H]1[C@@H](CO[P@](=O)(O[P@@](=O)(C(CCCC(P(... 6.0000 \n",
"54 CCCCCCSC[C@@H](C(=O)NCC(=O)O)NC(=O)CC[C@@H](C(... 10.0000 \n",
"55 c1ccccc1 175.0000 \n",
"... ... ... \n",
"105118 O[C@@H]([C@H](Cc1ccccc1)NC(=O)[C@H](C(C)C)NC(=... 0.0045 \n",
"105119 O[C@@H]([C@H](Cc1ccccc1)NC(=O)[C@H](C(C)C)NC(=... 0.0045 \n",
"105124 O[C@@H]1[C@@H](COP(=O)(O)O)O[C@H]([C@@H]1OP(=O... 125.0000 \n",
"105133 CC(C[C@@H](B(O)O)NC(=O)[C@@H]1CCCN1C(=O)[C@@H]... 2.0000 \n",
"105138 CC[Se]C(=N)N 0.0390 \n",
"\n",
"[13645 rows x 3 columns]"
]
},
"execution_count": 11,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df_biolip"
]
},
{
"cell_type": "code",
"execution_count": 12,
"id": "195f92db-fe06-4d03-8500-8d6c310a3347",
"metadata": {},
"outputs": [],
"source": [
"df_all = pd.concat([df_pdbbind,df_bindingdb,df_moad,df_biolip]).reset_index()"
]
},
{
"cell_type": "code",
"execution_count": 13,
"id": "d25c1e24-6566-4944-a0b4-944b3c8dbc6f",
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"674728"
]
},
"execution_count": 13,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"len(df_all)"
]
},
{
"cell_type": "code",
"execution_count": 14,
"id": "c8287da2-cfdf-4d89-b175-f4c6b38ff8ac",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"INFO: Pandarallel will run on 32 workers.\n",
"INFO: Pandarallel will use Memory file system to transfer data between the main process and workers.\n"
]
}
],
"source": [
"from pandarallel import pandarallel\n",
"pandarallel.initialize()"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "de5ffc4a-afb7-4a26-8d57-509c2278d750",
"metadata": {},
"outputs": [],
"source": [
"df_all['maccs'] = df_all['smiles'].parallel_apply(get_maccs)"
]
},
{
"cell_type": "code",
"execution_count": 16,
"id": "59a6706d-dab9-4ee0-8ef6-33537a3622a4",
"metadata": {},
"outputs": [],
"source": [
"df_all.to_parquet('data/all_maccs.parquet')"
]
},
{
"cell_type": "code",
"execution_count": 17,
"id": "4ccf2ee5-d369-4c0e-bb91-792765d661bf",
"metadata": {},
"outputs": [],
"source": [
"import numpy as np"
]
},
{
"cell_type": "code",
"execution_count": 18,
"id": "8a4bbb18-e62f-4774-ac6b-8a1be68204c1",
"metadata": {},
"outputs": [],
"source": [
"df_all = pd.read_parquet('data/all_maccs.parquet')\n",
"df_all = df_all.dropna().reset_index(drop=True)"
]
},
{
"cell_type": "code",
"execution_count": 19,
"id": "d210fe56-a7eb-4adc-a77a-14c0c6d0034e",
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"662484"
]
},
"execution_count": 19,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"len(df_all)"
]
},
{
"cell_type": "code",
"execution_count": 20,
"id": "d12b365d-98bd-4b61-b836-1a08d2e55418",
"metadata": {},
"outputs": [],
"source": [
"maccs = df_all['maccs'].to_numpy()\n",
"#df_reindex[df_reindex.duplicated(keep='first')].reset_index()"
]
},
{
"cell_type": "code",
"execution_count": 21,
"id": "80c15210-1af3-436e-970b-f81fc596fb41",
"metadata": {},
"outputs": [],
"source": [
"df_maccs = pd.DataFrame(np.vstack(maccs))"
]
},
{
"cell_type": "code",
"execution_count": 22,
"id": "30c314b8-8fe7-48ae-a2b8-149de1471b0c",
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"0 int64\n",
"1 int64\n",
"2 int64\n",
"3 int64\n",
"4 int64\n",
"5 int64\n",
"dtype: object"
]
},
"execution_count": 22,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df_maccs.dtypes"
]
},
{
"cell_type": "code",
"execution_count": 23,
"id": "70a0a820-4d0c-4472-af96-9c301c0ab204",
"metadata": {},
"outputs": [],
"source": [
"df_expand = pd.concat([df_all[['seq','smiles','affinity_uM']],df_maccs],axis=1)"
]
},
{
"cell_type": "code",
"execution_count": 24,
"id": "13d092fa-5625-40d0-b7ec-e3405ea20279",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" seq | \n",
" smiles | \n",
" affinity_uM | \n",
" 0 | \n",
" 1 | \n",
" 2 | \n",
" 3 | \n",
" 4 | \n",
" 5 | \n",
"
\n",
" \n",
" \n",
" \n",
" 0 | \n",
" SMENFQKVEKIGEGTYGVVYKARNKLTGEVVALKKIRLDTETEGVP... | \n",
" OCC[C@@H]1CCCCN1c1cc(NCC2=CC=CN(C2)O)n2c(n1)c(... | \n",
" 0.2100 | \n",
" 2147484672 | \n",
" 36176898 | \n",
" 850664773 | \n",
" 3978479102 | \n",
" 1599828989 | \n",
" 252 | \n",
"
\n",
" \n",
" 1 | \n",
" EFSEWFHNILEEAEIIDQRYPVKGMHVWMPHGFMIRKNTLKILRRI... | \n",
" O[C@@H]1[C@@H](COS(=O)(=O)NC(=O)[C@@H]2CCC[NH2... | \n",
" 0.0500 | \n",
" 0 | \n",
" 1858306115 | \n",
" 4223456596 | \n",
" 4018595822 | \n",
" 4282121085 | \n",
" 124 | \n",
"
\n",
" \n",
" 2 | \n",
" RGSHMEDFVRQCFNPMIVELAEKAMKEYGEDPKIETNKFAAICTHL... | \n",
" CCNC(=O)c1nc([nH]c(=O)c1O)[C@@H]1CCCN1C(=O)C | \n",
" 2.0000 | \n",
" 0 | \n",
" 33947650 | \n",
" 2041877824 | \n",
" 3782085608 | \n",
" 4290771792 | \n",
" 252 | \n",
"
\n",
" \n",
" 3 | \n",
" QISVRGLAGVENVTELKKNFNRHLHFTLVKDRNVATPRDYYFALAH... | \n",
" OC[C@H]1O[C@H](C[C@H]([C@@H]1O)F)n1ccc(nc1=O)N... | \n",
" 6550.0000 | \n",
" 0 | \n",
" 1107566598 | \n",
" 1755681856 | \n",
" 3846453088 | \n",
" 4293647263 | \n",
" 124 | \n",
"
\n",
" \n",
" 4 | \n",
" YELPEDPRWELPRDRLVLGKPLGEGQVVLAEAIGLDKDKPNRVTKV... | \n",
" C[N@@H+]1CC[N@H+](CC1)Cc1ccc(cc1C(F)(F)F)NC(=O... | \n",
" 0.0077 | \n",
" 4194304 | \n",
" 4857858 | \n",
" 515249472 | \n",
" 3969840044 | \n",
" 2061460183 | \n",
" 252 | \n",
"
\n",
" \n",
" ... | \n",
" ... | \n",
" ... | \n",
" ... | \n",
" ... | \n",
" ... | \n",
" ... | \n",
" ... | \n",
" ... | \n",
" ... | \n",
"
\n",
" \n",
" 662479 | \n",
" PQITLWQRPLVTIKIGGQLKEALLDTGADDTVLEEMNLPGRWKPKM... | \n",
" O[C@@H]([C@H](Cc1ccccc1)NC(=O)[C@H](C(C)C)NC(=... | \n",
" 0.0045 | \n",
" 65536 | \n",
" 393216 | \n",
" 964698368 | \n",
" 369403648 | \n",
" 4284858000 | \n",
" 252 | \n",
"
\n",
" \n",
" 662480 | \n",
" PQITLWQRPLVTIKIGGQLKEALLDTGADDTVLEEMNLPGRWKPKM... | \n",
" O[C@@H]([C@H](Cc1ccccc1)NC(=O)[C@H](C(C)C)NC(=... | \n",
" 0.0045 | \n",
" 65536 | \n",
" 393216 | \n",
" 964698368 | \n",
" 369403648 | \n",
" 4284858000 | \n",
" 252 | \n",
"
\n",
" \n",
" 662481 | \n",
" SKVVVPAQGKKITLQNGKLNVPENPIIPYIEGDGIGVDVTPAMLKV... | \n",
" O[C@@H]1[C@@H](COP(=O)(O)O)O[C@H]([C@@H]1OP(=O... | \n",
" 125.0000 | \n",
" 67108864 | \n",
" 1115688962 | \n",
" 1771869508 | \n",
" 4018431718 | \n",
" 3744193341 | \n",
" 124 | \n",
"
\n",
" \n",
" 662482 | \n",
" ANIVGGIEYSINNASLCSVGFSVTRGATKGFVTAGHCGTVNATARI... | \n",
" CC(C[C@@H](B(O)O)NC(=O)[C@@H]1CCCN1C(=O)[C@@H]... | \n",
" 2.0000 | \n",
" 2097152 | \n",
" 137216 | \n",
" 958148868 | \n",
" 1746307978 | \n",
" 2067783280 | \n",
" 204 | \n",
"
\n",
" \n",
" 662483 | \n",
" KFPRVKNWELGSITYDTLCAQSQQDGPCTPRRCLGSLVLPRKLQTR... | \n",
" CC[Se]C(=N)N | \n",
" 0.0390 | \n",
" 16 | \n",
" 6144 | \n",
" 537396736 | \n",
" 2170880 | \n",
" 1510015504 | \n",
" 192 | \n",
"
\n",
" \n",
"
\n",
"
662484 rows × 9 columns
\n",
"
"
],
"text/plain": [
" seq \\\n",
"0 SMENFQKVEKIGEGTYGVVYKARNKLTGEVVALKKIRLDTETEGVP... \n",
"1 EFSEWFHNILEEAEIIDQRYPVKGMHVWMPHGFMIRKNTLKILRRI... \n",
"2 RGSHMEDFVRQCFNPMIVELAEKAMKEYGEDPKIETNKFAAICTHL... \n",
"3 QISVRGLAGVENVTELKKNFNRHLHFTLVKDRNVATPRDYYFALAH... \n",
"4 YELPEDPRWELPRDRLVLGKPLGEGQVVLAEAIGLDKDKPNRVTKV... \n",
"... ... \n",
"662479 PQITLWQRPLVTIKIGGQLKEALLDTGADDTVLEEMNLPGRWKPKM... \n",
"662480 PQITLWQRPLVTIKIGGQLKEALLDTGADDTVLEEMNLPGRWKPKM... \n",
"662481 SKVVVPAQGKKITLQNGKLNVPENPIIPYIEGDGIGVDVTPAMLKV... \n",
"662482 ANIVGGIEYSINNASLCSVGFSVTRGATKGFVTAGHCGTVNATARI... \n",
"662483 KFPRVKNWELGSITYDTLCAQSQQDGPCTPRRCLGSLVLPRKLQTR... \n",
"\n",
" smiles affinity_uM \\\n",
"0 OCC[C@@H]1CCCCN1c1cc(NCC2=CC=CN(C2)O)n2c(n1)c(... 0.2100 \n",
"1 O[C@@H]1[C@@H](COS(=O)(=O)NC(=O)[C@@H]2CCC[NH2... 0.0500 \n",
"2 CCNC(=O)c1nc([nH]c(=O)c1O)[C@@H]1CCCN1C(=O)C 2.0000 \n",
"3 OC[C@H]1O[C@H](C[C@H]([C@@H]1O)F)n1ccc(nc1=O)N... 6550.0000 \n",
"4 C[N@@H+]1CC[N@H+](CC1)Cc1ccc(cc1C(F)(F)F)NC(=O... 0.0077 \n",
"... ... ... \n",
"662479 O[C@@H]([C@H](Cc1ccccc1)NC(=O)[C@H](C(C)C)NC(=... 0.0045 \n",
"662480 O[C@@H]([C@H](Cc1ccccc1)NC(=O)[C@H](C(C)C)NC(=... 0.0045 \n",
"662481 O[C@@H]1[C@@H](COP(=O)(O)O)O[C@H]([C@@H]1OP(=O... 125.0000 \n",
"662482 CC(C[C@@H](B(O)O)NC(=O)[C@@H]1CCCN1C(=O)[C@@H]... 2.0000 \n",
"662483 CC[Se]C(=N)N 0.0390 \n",
"\n",
" 0 1 2 3 4 5 \n",
"0 2147484672 36176898 850664773 3978479102 1599828989 252 \n",
"1 0 1858306115 4223456596 4018595822 4282121085 124 \n",
"2 0 33947650 2041877824 3782085608 4290771792 252 \n",
"3 0 1107566598 1755681856 3846453088 4293647263 124 \n",
"4 4194304 4857858 515249472 3969840044 2061460183 252 \n",
"... ... ... ... ... ... ... \n",
"662479 65536 393216 964698368 369403648 4284858000 252 \n",
"662480 65536 393216 964698368 369403648 4284858000 252 \n",
"662481 67108864 1115688962 1771869508 4018431718 3744193341 124 \n",
"662482 2097152 137216 958148868 1746307978 2067783280 204 \n",
"662483 16 6144 537396736 2170880 1510015504 192 \n",
"\n",
"[662484 rows x 9 columns]"
]
},
"execution_count": 24,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df_expand"
]
},
{
"cell_type": "code",
"execution_count": 25,
"id": "30f7fff7-3cfe-41c8-97c9-666f3e256222",
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"Index(['seq', 'smiles', 'affinity_uM', 0, 1, 2, 3, 4, 5], dtype='object')"
]
},
"execution_count": 25,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df_expand.columns"
]
},
{
"cell_type": "code",
"execution_count": 26,
"id": "16d2b26e-984f-4c71-af19-a3e711ed9ca2",
"metadata": {},
"outputs": [],
"source": [
"df_reindex = df_expand.set_index([0,1,2,3,4,5,'seq'])"
]
},
{
"cell_type": "code",
"execution_count": 27,
"id": "27fa2150-8152-444b-ba5b-24bea39fc098",
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"Index(['smiles', 'affinity_uM'], dtype='object')"
]
},
"execution_count": 27,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df_reindex.columns"
]
},
{
"cell_type": "code",
"execution_count": 28,
"id": "89edacbc-52f3-4a76-90b0-95273f5e53b3",
"metadata": {},
"outputs": [],
"source": [
"df_nr = df_reindex[~df_reindex.duplicated(keep='first')].reset_index()\n",
"df_nr = df_nr.drop(columns=[0,1,2,3,4,5])"
]
},
{
"cell_type": "code",
"execution_count": 68,
"id": "6a704c5e-68a6-418f-bcad-8688a13ca1d6",
"metadata": {},
"outputs": [],
"source": [
"# final sanity checks"
]
},
{
"cell_type": "code",
"execution_count": 30,
"id": "0cad3882-975d-4693-aad1-63ec26646bd0",
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"/ccs/proj/stf006/glaser/conda-envs/bio/lib/python3.9/site-packages/pandas/core/arraylike.py:358: RuntimeWarning: divide by zero encountered in log\n",
" result = getattr(ufunc, method)(*inputs, **kwargs)\n"
]
}
],
"source": [
"df_nr['neg_log10_affinity_M'] = 6-np.log(df_nr['affinity_uM'])/np.log(10)"
]
},
{
"cell_type": "code",
"execution_count": 31,
"id": "c200e29a-3f14-41f4-b620-ccce0eb0d5ce",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" seq | \n",
" smiles | \n",
" affinity_uM | \n",
" neg_log10_affinity_M | \n",
"
\n",
" \n",
" \n",
" \n",
" 0 | \n",
" SMENFQKVEKIGEGTYGVVYKARNKLTGEVVALKKIRLDTETEGVP... | \n",
" OCC[C@@H]1CCCCN1c1cc(NCC2=CC=CN(C2)O)n2c(n1)c(... | \n",
" 0.2100 | \n",
" 6.677781 | \n",
"
\n",
" \n",
" 1 | \n",
" EFSEWFHNILEEAEIIDQRYPVKGMHVWMPHGFMIRKNTLKILRRI... | \n",
" O[C@@H]1[C@@H](COS(=O)(=O)NC(=O)[C@@H]2CCC[NH2... | \n",
" 0.0500 | \n",
" 7.301030 | \n",
"
\n",
" \n",
" 2 | \n",
" RGSHMEDFVRQCFNPMIVELAEKAMKEYGEDPKIETNKFAAICTHL... | \n",
" CCNC(=O)c1nc([nH]c(=O)c1O)[C@@H]1CCCN1C(=O)C | \n",
" 2.0000 | \n",
" 5.698970 | \n",
"
\n",
" \n",
" 3 | \n",
" QISVRGLAGVENVTELKKNFNRHLHFTLVKDRNVATPRDYYFALAH... | \n",
" OC[C@H]1O[C@H](C[C@H]([C@@H]1O)F)n1ccc(nc1=O)N... | \n",
" 6550.0000 | \n",
" 2.183759 | \n",
"
\n",
" \n",
" 4 | \n",
" YELPEDPRWELPRDRLVLGKPLGEGQVVLAEAIGLDKDKPNRVTKV... | \n",
" C[N@@H+]1CC[N@H+](CC1)Cc1ccc(cc1C(F)(F)F)NC(=O... | \n",
" 0.0077 | \n",
" 8.113509 | \n",
"
\n",
" \n",
" ... | \n",
" ... | \n",
" ... | \n",
" ... | \n",
" ... | \n",
"
\n",
" \n",
" 488472 | \n",
" IVEGSDAEIGMSPWQVMLFRKSPQELLCGASLISDRWVLTAAHCLL... | \n",
" CC(C[C@@H](C(=O)N1C=CC[C@H]1C(=O)N)NC(=O)[C@@H... | \n",
" 8.0000 | \n",
" 5.096910 | \n",
"
\n",
" \n",
" 488473 | \n",
" PQITLWQRPLVTIKIGGQLKEALLDTGADDTVLEEMNLPGRWKPKM... | \n",
" O[C@@H]([C@H](Cc1ccccc1)NC(=O)[C@H](C(C)C)NC(=... | \n",
" 0.0045 | \n",
" 8.346787 | \n",
"
\n",
" \n",
" 488474 | \n",
" SKVVVPAQGKKITLQNGKLNVPENPIIPYIEGDGIGVDVTPAMLKV... | \n",
" O[C@@H]1[C@@H](COP(=O)(O)O)O[C@H]([C@@H]1OP(=O... | \n",
" 125.0000 | \n",
" 3.903090 | \n",
"
\n",
" \n",
" 488475 | \n",
" ANIVGGIEYSINNASLCSVGFSVTRGATKGFVTAGHCGTVNATARI... | \n",
" CC(C[C@@H](B(O)O)NC(=O)[C@@H]1CCCN1C(=O)[C@@H]... | \n",
" 2.0000 | \n",
" 5.698970 | \n",
"
\n",
" \n",
" 488476 | \n",
" KFPRVKNWELGSITYDTLCAQSQQDGPCTPRRCLGSLVLPRKLQTR... | \n",
" CC[Se]C(=N)N | \n",
" 0.0390 | \n",
" 7.408935 | \n",
"
\n",
" \n",
"
\n",
"
488477 rows × 4 columns
\n",
"
"
],
"text/plain": [
" seq \\\n",
"0 SMENFQKVEKIGEGTYGVVYKARNKLTGEVVALKKIRLDTETEGVP... \n",
"1 EFSEWFHNILEEAEIIDQRYPVKGMHVWMPHGFMIRKNTLKILRRI... \n",
"2 RGSHMEDFVRQCFNPMIVELAEKAMKEYGEDPKIETNKFAAICTHL... \n",
"3 QISVRGLAGVENVTELKKNFNRHLHFTLVKDRNVATPRDYYFALAH... \n",
"4 YELPEDPRWELPRDRLVLGKPLGEGQVVLAEAIGLDKDKPNRVTKV... \n",
"... ... \n",
"488472 IVEGSDAEIGMSPWQVMLFRKSPQELLCGASLISDRWVLTAAHCLL... \n",
"488473 PQITLWQRPLVTIKIGGQLKEALLDTGADDTVLEEMNLPGRWKPKM... \n",
"488474 SKVVVPAQGKKITLQNGKLNVPENPIIPYIEGDGIGVDVTPAMLKV... \n",
"488475 ANIVGGIEYSINNASLCSVGFSVTRGATKGFVTAGHCGTVNATARI... \n",
"488476 KFPRVKNWELGSITYDTLCAQSQQDGPCTPRRCLGSLVLPRKLQTR... \n",
"\n",
" smiles affinity_uM \\\n",
"0 OCC[C@@H]1CCCCN1c1cc(NCC2=CC=CN(C2)O)n2c(n1)c(... 0.2100 \n",
"1 O[C@@H]1[C@@H](COS(=O)(=O)NC(=O)[C@@H]2CCC[NH2... 0.0500 \n",
"2 CCNC(=O)c1nc([nH]c(=O)c1O)[C@@H]1CCCN1C(=O)C 2.0000 \n",
"3 OC[C@H]1O[C@H](C[C@H]([C@@H]1O)F)n1ccc(nc1=O)N... 6550.0000 \n",
"4 C[N@@H+]1CC[N@H+](CC1)Cc1ccc(cc1C(F)(F)F)NC(=O... 0.0077 \n",
"... ... ... \n",
"488472 CC(C[C@@H](C(=O)N1C=CC[C@H]1C(=O)N)NC(=O)[C@@H... 8.0000 \n",
"488473 O[C@@H]([C@H](Cc1ccccc1)NC(=O)[C@H](C(C)C)NC(=... 0.0045 \n",
"488474 O[C@@H]1[C@@H](COP(=O)(O)O)O[C@H]([C@@H]1OP(=O... 125.0000 \n",
"488475 CC(C[C@@H](B(O)O)NC(=O)[C@@H]1CCCN1C(=O)[C@@H]... 2.0000 \n",
"488476 CC[Se]C(=N)N 0.0390 \n",
"\n",
" neg_log10_affinity_M \n",
"0 6.677781 \n",
"1 7.301030 \n",
"2 5.698970 \n",
"3 2.183759 \n",
"4 8.113509 \n",
"... ... \n",
"488472 5.096910 \n",
"488473 8.346787 \n",
"488474 3.903090 \n",
"488475 5.698970 \n",
"488476 7.408935 \n",
"\n",
"[488477 rows x 4 columns]"
]
},
"execution_count": 31,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df_nr"
]
},
{
"cell_type": "code",
"execution_count": 32,
"id": "7f4027a2-0a5f-47bf-8a34-0c6a73b9b112",
"metadata": {},
"outputs": [],
"source": [
"df = df_nr[np.isfinite(df_nr['neg_log10_affinity_M'])]"
]
},
{
"cell_type": "code",
"execution_count": 33,
"id": "b4b9acd7-7784-492b-9fa3-b7fad9d18a9d",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"INFO: Pandarallel will run on 32 workers.\n",
"INFO: Pandarallel will use Memory file system to transfer data between the main process and workers.\n"
]
}
],
"source": [
"from pandarallel import pandarallel\n",
"pandarallel.initialize()\n"
]
},
{
"cell_type": "code",
"execution_count": 34,
"id": "eb99774f-9bcc-454d-b5e5-a8470223d6ca",
"metadata": {},
"outputs": [],
"source": [
"from rdkit import Chem\n",
"def make_canonical(smi):\n",
" try:\n",
" return Chem.MolToSmiles(Chem.MolFromSmiles(smi))\n",
" except:\n",
" return smi"
]
},
{
"cell_type": "code",
"execution_count": 35,
"id": "4d44bd8e-f2e1-44b4-aea7-40b4437baf44",
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
":1: SettingWithCopyWarning: \n",
"A value is trying to be set on a copy of a slice from a DataFrame.\n",
"Try using .loc[row_indexer,col_indexer] = value instead\n",
"\n",
"See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
" df['smiles_can'] = df['smiles'].parallel_apply(make_canonical)\n"
]
}
],
"source": [
"df['smiles_can'] = df['smiles'].parallel_apply(make_canonical)"
]
},
{
"cell_type": "code",
"execution_count": 36,
"id": "07ffdeb1-f4fa-4776-9fea-a18439e03d2e",
"metadata": {},
"outputs": [],
"source": [
"df = df[(df['neg_log10_affinity_M']>0) & (df['neg_log10_affinity_M']<15)].reset_index()"
]
},
{
"cell_type": "code",
"execution_count": 37,
"id": "8f949038-d07d-4d3a-a47e-b825cc9018ca",
"metadata": {},
"outputs": [],
"source": [
"from sklearn.preprocessing import StandardScaler"
]
},
{
"cell_type": "code",
"execution_count": 38,
"id": "0c027988-0b44-4010-ad61-7d70eead1654",
"metadata": {},
"outputs": [],
"source": [
"scaler = StandardScaler()"
]
},
{
"cell_type": "code",
"execution_count": 39,
"id": "6aeba020-b6ff-4633-902e-4df74463eb2f",
"metadata": {},
"outputs": [],
"source": [
"df['affinity'] = scaler.fit_transform(df['neg_log10_affinity_M'].values.reshape(-1,1))"
]
},
{
"cell_type": "code",
"execution_count": 40,
"id": "91196eee-5fd0-4aa4-927a-5c1a3f436ac8",
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(array([6.86202031]), array([2.57502859]))"
]
},
"execution_count": 40,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"scaler.mean_, scaler.var_"
]
},
{
"cell_type": "code",
"execution_count": 43,
"id": "56269dcb-e691-4759-949d-7bfdd02f5fd4",
"metadata": {},
"outputs": [],
"source": [
"df = df.drop(columns='index')"
]
},
{
"cell_type": "code",
"execution_count": 45,
"id": "c6c64066-4032-4247-a8b9-00388176cc7b",
"metadata": {},
"outputs": [],
"source": [
"df.to_parquet('data/all.parquet')"
]
},
{
"cell_type": "code",
"execution_count": 46,
"id": "469cf0dd-7b87-4245-973c-2a445e1fcca9",
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"Index(['seq', 'smiles', 'affinity_uM', 'neg_log10_affinity_M', 'smiles_can',\n",
" 'affinity'],\n",
" dtype='object')"
]
},
"execution_count": 46,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df.columns"
]
},
{
"cell_type": "code",
"execution_count": 47,
"id": "d91c0d91-474c-4ab2-9a5e-3b7861f7a832",
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
"