jglaser
commited on
Commit
·
1d7245f
0
Parent(s):
initial commit
Browse files- .gitattributes +1 -0
- .gitignore +1 -0
- README.md +34 -0
- batch_download.sh +114 -0
- data/pdb.parquet +3 -0
- get_pdb_ids.py +14 -0
- parse_complexes.py +196 -0
- pdb.slurm +10 -0
- pdb_protein_ligand_complexes.py +130 -0
- requirements.txt +5 -0
- rsync.sh +14 -0
- split_complex.py +123 -0
.gitattributes
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
data/pdb.parquet filter=lfs diff=lfs merge=lfs -text
|
.gitignore
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
slurm-*
|
README.md
ADDED
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
## How to use the data sets
|
2 |
+
|
3 |
+
This dataset contains more about 80,000 unique pairs of protein sequences and ligand SMILES, and the coordinates
|
4 |
+
of their complexes from the PDB. Only ligands with a molecular weight >= 100 Da are included.
|
5 |
+
|
6 |
+
SMILES are assumed to be tokenized by the regex from P. Schwaller.
|
7 |
+
|
8 |
+
Every (x,y,z) ligand coordinate maps onto a SMILES token, and is *nan* if the token does not represent an atom
|
9 |
+
|
10 |
+
Every receptor coordinate maps onto the Calpha coordinate of that residue.
|
11 |
+
|
12 |
+
The dataset can be used to fine-tune a language model, all data comes from PDBind-cn.
|
13 |
+
|
14 |
+
### Use the already preprocessed data
|
15 |
+
|
16 |
+
Load a test/train split using
|
17 |
+
|
18 |
+
```
|
19 |
+
from datasets import load_dataset
|
20 |
+
train = load_dataset("jglaser/pdb_protein_ligand_complexes",split='train[:90%]')
|
21 |
+
validation = load_dataset("jglaser/pdb_protein_ligand_complexes",split='train[90%:]')
|
22 |
+
```
|
23 |
+
|
24 |
+
### Manual update from PDB
|
25 |
+
|
26 |
+
```
|
27 |
+
# download the PDB archive into folder pdb/
|
28 |
+
sh rsync.sh 24 # number of parallel download processes
|
29 |
+
|
30 |
+
# extract sequences and coordinates in parallel
|
31 |
+
sbatch pdb.slurm
|
32 |
+
# or
|
33 |
+
mpirun -n 42 parse_complexes.py # desired number of tasks
|
34 |
+
```
|
batch_download.sh
ADDED
@@ -0,0 +1,114 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
#!/bin/bash
|
2 |
+
|
3 |
+
# Script to download files from RCSB http file download services.
|
4 |
+
# Use the -h switch to get help on usage.
|
5 |
+
|
6 |
+
if ! command -v curl &> /dev/null
|
7 |
+
then
|
8 |
+
echo "'curl' could not be found. You need to install 'curl' for this script to work."
|
9 |
+
exit 1
|
10 |
+
fi
|
11 |
+
|
12 |
+
PROGNAME=$0
|
13 |
+
BASE_URL="https://files.rcsb.org/download"
|
14 |
+
|
15 |
+
usage() {
|
16 |
+
cat << EOF >&2
|
17 |
+
Usage: $PROGNAME -f <file> [-o <dir>] [-c] [-p]
|
18 |
+
|
19 |
+
-f <file>: the input file containing a comma-separated list of PDB ids
|
20 |
+
-o <dir>: the output dir, default: current dir
|
21 |
+
-c : download a cif.gz file for each PDB id
|
22 |
+
-p : download a pdb.gz file for each PDB id (not available for large structures)
|
23 |
+
-a : download a pdb1.gz file (1st bioassembly) for each PDB id (not available for large structures)
|
24 |
+
-x : download a xml.gz file for each PDB id
|
25 |
+
-s : download a sf.cif.gz file for each PDB id (diffraction only)
|
26 |
+
-m : download a mr.gz file for each PDB id (NMR only)
|
27 |
+
-r : download a mr.str.gz for each PDB id (NMR only)
|
28 |
+
EOF
|
29 |
+
exit 1
|
30 |
+
}
|
31 |
+
|
32 |
+
download() {
|
33 |
+
url="$BASE_URL/$1"
|
34 |
+
out=$2/$1
|
35 |
+
echo "Downloading $url to $out"
|
36 |
+
curl -s -f $url -o $out || echo "Failed to download $url"
|
37 |
+
}
|
38 |
+
|
39 |
+
listfile=""
|
40 |
+
outdir="."
|
41 |
+
cif=false
|
42 |
+
pdb=false
|
43 |
+
pdb1=false
|
44 |
+
xml=false
|
45 |
+
sf=false
|
46 |
+
mr=false
|
47 |
+
mrstr=false
|
48 |
+
while getopts f:o:cpaxsmr o
|
49 |
+
do
|
50 |
+
case $o in
|
51 |
+
(f) listfile=$OPTARG;;
|
52 |
+
(o) outdir=$OPTARG;;
|
53 |
+
(c) cif=true;;
|
54 |
+
(p) pdb=true;;
|
55 |
+
(a) pdb1=true;;
|
56 |
+
(x) xml=true;;
|
57 |
+
(s) sf=true;;
|
58 |
+
(m) mr=true;;
|
59 |
+
(r) mrstr=true;;
|
60 |
+
(*) usage
|
61 |
+
esac
|
62 |
+
done
|
63 |
+
shift "$((OPTIND - 1))"
|
64 |
+
|
65 |
+
if [ "$listfile" == "" ]
|
66 |
+
then
|
67 |
+
echo "Parameter -f must be provided"
|
68 |
+
exit 1
|
69 |
+
fi
|
70 |
+
contents=$(cat $listfile)
|
71 |
+
|
72 |
+
# see https://stackoverflow.com/questions/918886/how-do-i-split-a-string-on-a-delimiter-in-bash#tab-top
|
73 |
+
IFS=',' read -ra tokens <<< "$contents"
|
74 |
+
|
75 |
+
for token in "${tokens[@]}"
|
76 |
+
do
|
77 |
+
if [ "$cif" == true ]
|
78 |
+
then
|
79 |
+
download ${token}.cif.gz $outdir
|
80 |
+
fi
|
81 |
+
if [ "$pdb" == true ]
|
82 |
+
then
|
83 |
+
download ${token}.pdb.gz $outdir
|
84 |
+
fi
|
85 |
+
if [ "$pdb1" == true ]
|
86 |
+
then
|
87 |
+
download ${token}.pdb1.gz $outdir
|
88 |
+
fi
|
89 |
+
if [ "$xml" == true ]
|
90 |
+
then
|
91 |
+
download ${token}.xml.gz $outdir
|
92 |
+
fi
|
93 |
+
if [ "$sf" == true ]
|
94 |
+
then
|
95 |
+
download ${token}-sf.cif.gz $outdir
|
96 |
+
fi
|
97 |
+
if [ "$mr" == true ]
|
98 |
+
then
|
99 |
+
download ${token}.mr.gz $outdir
|
100 |
+
fi
|
101 |
+
if [ "$mrstr" == true ]
|
102 |
+
then
|
103 |
+
download ${token}_mr.str.gz $outdir
|
104 |
+
fi
|
105 |
+
|
106 |
+
done
|
107 |
+
|
108 |
+
|
109 |
+
|
110 |
+
|
111 |
+
|
112 |
+
|
113 |
+
|
114 |
+
|
data/pdb.parquet
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c503ec9218e87f192d710838a0123f01ac1e48b75b288f0adc229cb025d1b592
|
3 |
+
size 1005021071
|
get_pdb_ids.py
ADDED
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from rcsbsearch import Terminal
|
2 |
+
from rcsbsearch import rcsb_attributes as attrs
|
3 |
+
|
4 |
+
# Create terminals for each query
|
5 |
+
q1 = attrs.rcsb_entry_info.nonpolymer_entity_count > 0
|
6 |
+
q2 = attrs.rcsb_entry_info.polymer_entity_count_protein > 0
|
7 |
+
q3 = Terminal('chem_comp.formula_weight','greater_or_equal',150,service='text_chem')
|
8 |
+
|
9 |
+
# combined using bitwise operators (&, |, ~, etc)
|
10 |
+
query = q1 & q2 & q3 # AND of all queries
|
11 |
+
|
12 |
+
# Call the query to execute it
|
13 |
+
for assemblyid in query("entry"):
|
14 |
+
print(assemblyid)
|
parse_complexes.py
ADDED
@@ -0,0 +1,196 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
#!/usr/bin/env python
|
2 |
+
|
3 |
+
# Split a protein-ligand complex into protein and ligands and assign ligand bond orders using SMILES strings from Ligand Export
|
4 |
+
# Code requires Python 3.6
|
5 |
+
|
6 |
+
import sys
|
7 |
+
from prody import *
|
8 |
+
import pandas as pd
|
9 |
+
from rdkit import Chem
|
10 |
+
from rdkit.Chem import AllChem
|
11 |
+
from io import StringIO
|
12 |
+
import requests
|
13 |
+
|
14 |
+
from mpi4py import MPI
|
15 |
+
from mpi4py.futures import MPICommExecutor
|
16 |
+
from mpi4py.futures import MPIPoolExecutor
|
17 |
+
|
18 |
+
import re
|
19 |
+
from functools import partial
|
20 |
+
import gzip
|
21 |
+
from rdkit.Chem.Descriptors import ExactMolWt
|
22 |
+
import numpy as np
|
23 |
+
|
24 |
+
import os
|
25 |
+
|
26 |
+
# minimum molecular weight to consider sth a ligand
|
27 |
+
mol_wt_cutoff = 100
|
28 |
+
|
29 |
+
# all punctuation
|
30 |
+
punctuation_regex = r"""(\(|\)|\.|=|#|-|\+|\\|\/|:|~|@|\?|>>?|\*|\$|\%[0-9]{2}|[0-9])"""
|
31 |
+
|
32 |
+
# tokenization regex (Schwaller)
|
33 |
+
molecule_regex = r"""(\[[^\]]+]|Br?|Cl?|N|O|S|P|F|I|b|c|n|o|s|p|\(|\)|\.|=|#|-|\+|\\|\/|:|~|@|\?|>>?|\*|\$|\%[0-9]{2}|[0-9])"""
|
34 |
+
|
35 |
+
max_seq = 2046 # = 2048 - 2 (accounting for [CLS] and [SEP])
|
36 |
+
max_smiles = 510 # = 512 - 2
|
37 |
+
|
38 |
+
def get_protein_sequence_and_coords(receptor):
|
39 |
+
calpha = receptor.select('calpha')
|
40 |
+
xyz = calpha.getCoords()
|
41 |
+
seq = calpha.getSequence()
|
42 |
+
return seq, xyz.tolist()
|
43 |
+
|
44 |
+
def tokenize_ligand(mol):
|
45 |
+
# convert to SMILES and map atoms
|
46 |
+
smi = Chem.MolToSmiles(mol)
|
47 |
+
|
48 |
+
# position of atoms in SMILES (not counting punctuation)
|
49 |
+
atom_order = [int(s) for s in list(filter(None,re.sub(r'[\[\]]','',mol.GetProp("_smilesAtomOutputOrder")).split(',')))]
|
50 |
+
|
51 |
+
# tokenize the SMILES
|
52 |
+
tokens = list(filter(None, re.split(molecule_regex, smi)))
|
53 |
+
|
54 |
+
# remove punctuation
|
55 |
+
masked_tokens = [re.sub(punctuation_regex,'',s) for s in tokens]
|
56 |
+
|
57 |
+
k = 0
|
58 |
+
token_pos = []
|
59 |
+
for i,token in enumerate(masked_tokens):
|
60 |
+
if token != '':
|
61 |
+
token_pos.append(tuple(mol.GetConformer().GetAtomPosition(atom_order[k])))
|
62 |
+
k += 1
|
63 |
+
else:
|
64 |
+
token_pos.append((np.nan, np.nan, np.nan))
|
65 |
+
|
66 |
+
return smi, token_pos
|
67 |
+
|
68 |
+
def read_ligand_expo():
|
69 |
+
"""
|
70 |
+
Read Ligand Expo data, try to find a file called
|
71 |
+
Components-smiles-stereo-oe.smi in the current directory.
|
72 |
+
If you can't find the file, grab it from the RCSB
|
73 |
+
:return: Ligand Expo as a dictionary with ligand id as the key
|
74 |
+
"""
|
75 |
+
file_name = "Components-smiles-stereo-oe.smi"
|
76 |
+
try:
|
77 |
+
df = pd.read_csv(file_name, sep="\t",
|
78 |
+
header=None,
|
79 |
+
names=["SMILES", "ID", "Name"])
|
80 |
+
except FileNotFoundError:
|
81 |
+
url = f"http://ligand-expo.rcsb.org/dictionaries/{file_name}"
|
82 |
+
print(url)
|
83 |
+
r = requests.get(url, allow_redirects=True)
|
84 |
+
open('Components-smiles-stereo-oe.smi', 'wb').write(r.content)
|
85 |
+
df = pd.read_csv(file_name, sep="\t",
|
86 |
+
header=None,
|
87 |
+
names=["SMILES", "ID", "Name"])
|
88 |
+
df.set_index("ID", inplace=True)
|
89 |
+
return df.to_dict()
|
90 |
+
|
91 |
+
|
92 |
+
def get_pdb_components(pdb_id):
|
93 |
+
"""
|
94 |
+
Split a protein-ligand pdb into protein and ligand components
|
95 |
+
:param pdb_id:
|
96 |
+
:return:
|
97 |
+
"""
|
98 |
+
with gzip.open(pdb_id,'rt') as f:
|
99 |
+
pdb = parsePDBStream(f)
|
100 |
+
|
101 |
+
protein = pdb.select('protein')
|
102 |
+
ligand = pdb.select('not protein and not water')
|
103 |
+
return protein, ligand
|
104 |
+
|
105 |
+
|
106 |
+
def process_ligand(ligand, res_name, expo_dict):
|
107 |
+
"""
|
108 |
+
Add bond orders to a pdb ligand
|
109 |
+
1. Select the ligand component with name "res_name"
|
110 |
+
2. Get the corresponding SMILES from the Ligand Expo dictionary
|
111 |
+
3. Create a template molecule from the SMILES in step 2
|
112 |
+
4. Write the PDB file to a stream
|
113 |
+
5. Read the stream into an RDKit molecule
|
114 |
+
6. Assign the bond orders from the template from step 3
|
115 |
+
:param ligand: ligand as generated by prody
|
116 |
+
:param res_name: residue name of ligand to extract
|
117 |
+
:param expo_dict: dictionary with LigandExpo
|
118 |
+
:return: molecule with bond orders assigned
|
119 |
+
"""
|
120 |
+
output = StringIO()
|
121 |
+
sub_mol = ligand.select(f"resname {res_name}")
|
122 |
+
sub_smiles = expo_dict['SMILES'][res_name]
|
123 |
+
template = AllChem.MolFromSmiles(sub_smiles)
|
124 |
+
writePDBStream(output, sub_mol)
|
125 |
+
pdb_string = output.getvalue()
|
126 |
+
rd_mol = AllChem.MolFromPDBBlock(pdb_string)
|
127 |
+
new_mol = AllChem.AssignBondOrdersFromTemplate(template, rd_mol)
|
128 |
+
return new_mol, template
|
129 |
+
|
130 |
+
def process_entry(df_dict, pdb_fn):
|
131 |
+
try:
|
132 |
+
"""
|
133 |
+
Slit pdb into protein and ligands,
|
134 |
+
parse protein sequence and ligand tokens
|
135 |
+
:param df_dict: ligand expo data
|
136 |
+
:param pdb_fn: pdb entry file name
|
137 |
+
:return:
|
138 |
+
"""
|
139 |
+
protein, ligand = get_pdb_components(pdb_fn)
|
140 |
+
|
141 |
+
ligand_mols = []
|
142 |
+
ligand_names = []
|
143 |
+
|
144 |
+
if ligand is not None:
|
145 |
+
# filter ligands by molecular weight
|
146 |
+
res_name_list = list(set(ligand.getResnames()))
|
147 |
+
for res in res_name_list:
|
148 |
+
mol, template = process_ligand(ligand, res, df_dict)
|
149 |
+
|
150 |
+
mol_wt = ExactMolWt(template)
|
151 |
+
|
152 |
+
if mol_wt >= mol_wt_cutoff:
|
153 |
+
ligand_mols.append(mol)
|
154 |
+
ligand_names.append(res)
|
155 |
+
|
156 |
+
ligand_smiles = []
|
157 |
+
ligand_xyz = []
|
158 |
+
|
159 |
+
pdb_name = os.path.basename(pdb_fn).split('.')[-3][3:]
|
160 |
+
for mol, name in zip(ligand_mols, ligand_names):
|
161 |
+
print('Processing {} and {}'.format(pdb_name, name))
|
162 |
+
smi, xyz = tokenize_ligand(mol)
|
163 |
+
ligand_smiles.append(smi)
|
164 |
+
ligand_xyz.append(xyz)
|
165 |
+
|
166 |
+
seq, receptor_xyz = get_protein_sequence_and_coords(protein)
|
167 |
+
return pdb_name, seq, receptor_xyz, ligand_names, ligand_smiles, ligand_xyz
|
168 |
+
except Exception as e:
|
169 |
+
print(repr(e))
|
170 |
+
|
171 |
+
if __name__ == '__main__':
|
172 |
+
import glob
|
173 |
+
|
174 |
+
filenames = glob.glob('pdb/*/*.gz')
|
175 |
+
filenames = sorted(filenames)
|
176 |
+
comm = MPI.COMM_WORLD
|
177 |
+
with MPICommExecutor(comm, root=0) as executor:
|
178 |
+
# with MPIPoolExecutor() as executor:
|
179 |
+
if executor is not None:
|
180 |
+
# read ligand table
|
181 |
+
df_dict = read_ligand_expo()
|
182 |
+
|
183 |
+
result = executor.map(partial(process_entry, df_dict), filenames, chunksize=512)
|
184 |
+
result = list(result)
|
185 |
+
|
186 |
+
# expand sequences and ligands
|
187 |
+
pdb_id = [r[0] for r in result if r is not None for ligand in r[3]]
|
188 |
+
seq = [r[1] for r in result if r is not None for ligand in r[3]]
|
189 |
+
receptor_xyz = [r[2] for r in result if r is not None for ligand in r[3]]
|
190 |
+
lig_id = [l for r in result if r is not None for l in r[3]]
|
191 |
+
lig_smiles = [s for r in result if r is not None for s in r[4]]
|
192 |
+
lig_xyz = [xyz for r in result if r is not None for xyz in r[5]]
|
193 |
+
|
194 |
+
import pandas as pd
|
195 |
+
df = pd.DataFrame({'pdb_id': pdb_id, 'lig_id': lig_id, 'seq': seq, 'smiles': lig_smiles, 'receptor_xyz': receptor_xyz, 'ligand_xyz': lig_xyz})
|
196 |
+
df.to_parquet('data/pdb.parquet',index=False)
|
pdb.slurm
ADDED
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
#!/bin/bash
|
2 |
+
#SBATCH -J preprocess
|
3 |
+
#SBATCH -p batch
|
4 |
+
#SBATCH -A STF006
|
5 |
+
#SBATCH -t 3:00:00
|
6 |
+
#SBATCH -N 36
|
7 |
+
#SBATCH --ntasks-per-node=16
|
8 |
+
|
9 |
+
export PYTHONUNBUFFERED=1
|
10 |
+
srun python parse_complexes.py
|
pdb_protein_ligand_complexes.py
ADDED
@@ -0,0 +1,130 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# coding=utf-8
|
2 |
+
# Copyright 2020 The HuggingFace Datasets Authors and the current dataset script contributor.
|
3 |
+
#
|
4 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
5 |
+
# you may not use this file except in compliance with the License.
|
6 |
+
# You may obtain a copy of the License at
|
7 |
+
#
|
8 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
9 |
+
#
|
10 |
+
# Unless required by applicable law or agreed to in writing, software
|
11 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
12 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
13 |
+
# See the License for the specific language governing permissions and
|
14 |
+
# limitations under the License.
|
15 |
+
"""TODO: A dataset of protein sequences, ligand SMILES, and complex coordinates."""
|
16 |
+
|
17 |
+
import huggingface_hub
|
18 |
+
import os
|
19 |
+
import pyarrow.parquet as pq
|
20 |
+
import datasets
|
21 |
+
|
22 |
+
|
23 |
+
# TODO: Add BibTeX citation
|
24 |
+
# Find for instance the citation on arxiv or on the dataset repo/website
|
25 |
+
_CITATION = """\
|
26 |
+
@InProceedings{huggingface:dataset,
|
27 |
+
title = {jglaser/pdb_protein_ligand_complexes},
|
28 |
+
author={Jens Glaser, ORNL
|
29 |
+
},
|
30 |
+
year={2022}
|
31 |
+
}
|
32 |
+
"""
|
33 |
+
|
34 |
+
# TODO: Add description of the dataset here
|
35 |
+
# You can copy an official description
|
36 |
+
_DESCRIPTION = """\
|
37 |
+
A dataset to fine-tune language models on protein-ligand complex structures
|
38 |
+
"""
|
39 |
+
|
40 |
+
# TODO: Add a link to an official homepage for the dataset here
|
41 |
+
_HOMEPAGE = ""
|
42 |
+
|
43 |
+
# TODO: Add the licence for the dataset here if you can find it
|
44 |
+
_LICENSE = "BSD two-clause"
|
45 |
+
|
46 |
+
# TODO: Add link to the official dataset URLs here
|
47 |
+
# The HuggingFace dataset library don't host the datasets but only point to the original files
|
48 |
+
# This can be an arbitrary nested dict/list of URLs (see below in `_split_generators` method)
|
49 |
+
_URL = "https://huggingface.co/datasets/jglaser/pdb_protein_ligand_complexes/resolve/main/"
|
50 |
+
_data_dir = "data/"
|
51 |
+
_file_names = {'default': _data_dir+'pdb.parquet'}
|
52 |
+
|
53 |
+
_URLs = {name: _URL+_file_names[name] for name in _file_names}
|
54 |
+
|
55 |
+
|
56 |
+
# TODO: Name of the dataset usually match the script name with CamelCase instead of snake_case
|
57 |
+
class ProteinLigandContacts(datasets.ArrowBasedBuilder):
|
58 |
+
"""List of protein sequences, ligand SMILES, and complex contacts."""
|
59 |
+
|
60 |
+
VERSION = datasets.Version("1.0.0")
|
61 |
+
|
62 |
+
def _info(self):
|
63 |
+
# TODO: This method specifies the datasets.DatasetInfo object which contains informations and typings for the dataset
|
64 |
+
#if self.config.name == "first_domain": # This is the name of the configuration selected in BUILDER_CONFIGS above
|
65 |
+
# features = datasets.Features(
|
66 |
+
# {
|
67 |
+
# "sentence": datasets.Value("string"),
|
68 |
+
# "option1": datasets.Value("string"),
|
69 |
+
# "answer": datasets.Value("string")
|
70 |
+
# # These are the features of your dataset like images, labels ...
|
71 |
+
# }
|
72 |
+
# )
|
73 |
+
#else: # This is an example to show how to have different features for "first_domain" and "second_domain"
|
74 |
+
features = datasets.Features(
|
75 |
+
{
|
76 |
+
"pdb_id": datasets.Value("string"),
|
77 |
+
"lig_id": datasets.Value("string"),
|
78 |
+
"seq": datasets.Value("string"),
|
79 |
+
"smiles": datasets.Value("string"),
|
80 |
+
"ligand_xyz": datasets.Sequence(datasets.Sequence(datasets.Value('float32'))),
|
81 |
+
"receptor_xyz": datasets.Sequence(datasets.Sequence(datasets.Value('float32'))),
|
82 |
+
# These are the features of your dataset like images, labels ...
|
83 |
+
}
|
84 |
+
)
|
85 |
+
return datasets.DatasetInfo(
|
86 |
+
# This is the description that will appear on the datasets page.
|
87 |
+
description=_DESCRIPTION,
|
88 |
+
# This defines the different columns of the dataset and their types
|
89 |
+
features=features, # Here we define them above because they are different between the two configurations
|
90 |
+
# If there's a common (input, target) tuple from the features,
|
91 |
+
# specify them here. They'll be used if as_supervised=True in
|
92 |
+
# builder.as_dataset.
|
93 |
+
supervised_keys=None,
|
94 |
+
# Homepage of the dataset for documentation
|
95 |
+
homepage=_HOMEPAGE,
|
96 |
+
# License for the dataset if available
|
97 |
+
license=_LICENSE,
|
98 |
+
# Citation for the dataset
|
99 |
+
citation=_CITATION,
|
100 |
+
)
|
101 |
+
|
102 |
+
def _split_generators(self, dl_manager):
|
103 |
+
"""Returns SplitGenerators."""
|
104 |
+
# TODO: This method is tasked with downloading/extracting the data and defining the splits depending on the configuration
|
105 |
+
# If several configurations are possible (listed in BUILDER_CONFIGS), the configuration selected by the user is in self.config.name
|
106 |
+
|
107 |
+
# dl_manager is a datasets.download.DownloadManager that can be used to download and extract URLs
|
108 |
+
# It can accept any type or nested list/dict and will give back the same structure with the url replaced with path to local files.
|
109 |
+
# By default the archives will be extracted and a path to a cached folder where they are extracted is returned instead of the archive
|
110 |
+
files = dl_manager.download_and_extract(_URLs)
|
111 |
+
|
112 |
+
return [
|
113 |
+
datasets.SplitGenerator(
|
114 |
+
# These kwargs will be passed to _generate_examples
|
115 |
+
name=datasets.Split.TRAIN,
|
116 |
+
gen_kwargs={
|
117 |
+
'filepath': files["default"],
|
118 |
+
},
|
119 |
+
),
|
120 |
+
|
121 |
+
]
|
122 |
+
|
123 |
+
def _generate_tables(
|
124 |
+
self, filepath
|
125 |
+
):
|
126 |
+
from pyarrow import fs
|
127 |
+
local = fs.LocalFileSystem()
|
128 |
+
|
129 |
+
for i, f in enumerate([filepath]):
|
130 |
+
yield i, pq.read_table(f,filesystem=local)
|
requirements.txt
ADDED
@@ -0,0 +1,5 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
prody
|
2 |
+
mpi4py
|
3 |
+
numpy
|
4 |
+
pandas
|
5 |
+
rdkit
|
rsync.sh
ADDED
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
#!/bin/sh
|
2 |
+
|
3 |
+
NPROCS=$1
|
4 |
+
MIRRORDIR=pdb # your top level rsync directory
|
5 |
+
RSYNC=rsync # location of local rsync
|
6 |
+
|
7 |
+
SERVER=rsync.wwpdb.org::ftp # RCSB PDB server name
|
8 |
+
PORT=33444 # port RCSB PDB server is using
|
9 |
+
|
10 |
+
# get file list, remove first and last 3 lines from output (double-check that these lines are not needed)
|
11 |
+
${RSYNC} -lpt -v -z --delete --port=$PORT --no-h --list-only ${SERVER}/data/structures/divided/pdb/ | cut -c 44- | head -n -3 | tail -n +3 > dirlist.txt
|
12 |
+
|
13 |
+
cat dirlist.txt | xargs -n1 -P${NPROCS} -I% rsync -rlpt -v -z --port=$PORT -P ${SERVER}/data/structures/divided/pdb/% $MIRRORDIR
|
14 |
+
rm dirlist.txt
|
split_complex.py
ADDED
@@ -0,0 +1,123 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
#!/usr/bin/env python
|
2 |
+
|
3 |
+
# Split a protein-ligand complex into protein and ligands and assign ligand bond orders using SMILES strings from Ligand Export
|
4 |
+
# Code requires Python 3.6
|
5 |
+
|
6 |
+
import sys
|
7 |
+
from prody import *
|
8 |
+
import pandas as pd
|
9 |
+
from rdkit import Chem
|
10 |
+
from rdkit.Chem import AllChem
|
11 |
+
from io import StringIO
|
12 |
+
import requests
|
13 |
+
|
14 |
+
|
15 |
+
def read_ligand_expo():
|
16 |
+
"""
|
17 |
+
Read Ligand Expo data, try to find a file called
|
18 |
+
Components-smiles-stereo-oe.smi in the current directory.
|
19 |
+
If you can't find the file, grab it from the RCSB
|
20 |
+
:return: Ligand Expo as a dictionary with ligand id as the key
|
21 |
+
"""
|
22 |
+
file_name = "Components-smiles-stereo-oe.smi"
|
23 |
+
try:
|
24 |
+
df = pd.read_csv(file_name, sep="\t",
|
25 |
+
header=None,
|
26 |
+
names=["SMILES", "ID", "Name"])
|
27 |
+
except FileNotFoundError:
|
28 |
+
url = f"http://ligand-expo.rcsb.org/dictionaries/{file_name}"
|
29 |
+
print(url)
|
30 |
+
r = requests.get(url, allow_redirects=True)
|
31 |
+
open('Components-smiles-stereo-oe.smi', 'wb').write(r.content)
|
32 |
+
df = pd.read_csv(file_name, sep="\t",
|
33 |
+
header=None,
|
34 |
+
names=["SMILES", "ID", "Name"])
|
35 |
+
df.set_index("ID", inplace=True)
|
36 |
+
return df.to_dict()
|
37 |
+
|
38 |
+
|
39 |
+
def get_pdb_components(pdb_id):
|
40 |
+
"""
|
41 |
+
Split a protein-ligand pdb into protein and ligand components
|
42 |
+
:param pdb_id:
|
43 |
+
:return:
|
44 |
+
"""
|
45 |
+
pdb = parsePDB(pdb_id)
|
46 |
+
protein = pdb.select('protein')
|
47 |
+
ligand = pdb.select('not protein and not water')
|
48 |
+
return protein, ligand
|
49 |
+
|
50 |
+
|
51 |
+
def process_ligand(ligand, res_name, expo_dict):
|
52 |
+
"""
|
53 |
+
Add bond orders to a pdb ligand
|
54 |
+
1. Select the ligand component with name "res_name"
|
55 |
+
2. Get the corresponding SMILES from the Ligand Expo dictionary
|
56 |
+
3. Create a template molecule from the SMILES in step 2
|
57 |
+
4. Write the PDB file to a stream
|
58 |
+
5. Read the stream into an RDKit molecule
|
59 |
+
6. Assign the bond orders from the template from step 3
|
60 |
+
:param ligand: ligand as generated by prody
|
61 |
+
:param res_name: residue name of ligand to extract
|
62 |
+
:param expo_dict: dictionary with LigandExpo
|
63 |
+
:return: molecule with bond orders assigned
|
64 |
+
"""
|
65 |
+
output = StringIO()
|
66 |
+
sub_mol = ligand.select(f"resname {res_name}")
|
67 |
+
sub_smiles = expo_dict['SMILES'][res_name]
|
68 |
+
template = AllChem.MolFromSmiles(sub_smiles)
|
69 |
+
writePDBStream(output, sub_mol)
|
70 |
+
pdb_string = output.getvalue()
|
71 |
+
rd_mol = AllChem.MolFromPDBBlock(pdb_string)
|
72 |
+
new_mol = AllChem.AssignBondOrdersFromTemplate(template, rd_mol)
|
73 |
+
return new_mol
|
74 |
+
|
75 |
+
|
76 |
+
def write_pdb(protein, pdb_name):
|
77 |
+
"""
|
78 |
+
Write a prody protein to a pdb file
|
79 |
+
:param protein: protein object from prody
|
80 |
+
:param pdb_name: base name for the pdb file
|
81 |
+
:return: None
|
82 |
+
"""
|
83 |
+
output_pdb_name = f"{pdb_name}_protein.pdb"
|
84 |
+
writePDB(f"{output_pdb_name}", protein)
|
85 |
+
print(f"wrote {output_pdb_name}")
|
86 |
+
|
87 |
+
|
88 |
+
def write_sdf(new_mol, pdb_name, res_name):
|
89 |
+
"""
|
90 |
+
Write an RDKit molecule to an SD file
|
91 |
+
:param new_mol:
|
92 |
+
:param pdb_name:
|
93 |
+
:param res_name:
|
94 |
+
:return:
|
95 |
+
"""
|
96 |
+
outfile_name = f"{pdb_name}_{res_name}_ligand.sdf"
|
97 |
+
writer = Chem.SDWriter(f"{outfile_name}")
|
98 |
+
writer.write(new_mol)
|
99 |
+
print(f"wrote {outfile_name}")
|
100 |
+
|
101 |
+
|
102 |
+
def main(pdb_name):
|
103 |
+
"""
|
104 |
+
Read Ligand Expo data, split pdb into protein and ligands,
|
105 |
+
write protein pdb, write ligand sdf files
|
106 |
+
:param pdb_name: id from the pdb, doesn't need to have an extension
|
107 |
+
:return:
|
108 |
+
"""
|
109 |
+
df_dict = read_ligand_expo()
|
110 |
+
protein, ligand = get_pdb_components(pdb_name)
|
111 |
+
write_pdb(protein, pdb_name)
|
112 |
+
|
113 |
+
res_name_list = list(set(ligand.getResnames()))
|
114 |
+
for res in res_name_list:
|
115 |
+
new_mol = process_ligand(ligand, res, df_dict)
|
116 |
+
write_sdf(new_mol, pdb_name, res)
|
117 |
+
|
118 |
+
|
119 |
+
if __name__ == "__main__":
|
120 |
+
if len(sys.argv) == 2:
|
121 |
+
main(sys.argv[1])
|
122 |
+
else:
|
123 |
+
print("Usage: {sys.argv[1]} pdb_id", file=sys.stderr)
|