def _lookup(self, file_path):
# Assume one ligand in the pdbbind ligand SDF file
ligand = fo.read_sdf_to_mol(file_path,
sanitize=False,
add_hs=False,
remove_hs=True)[0]
return Chem.MolToSmiles(ligand)
def __getitem__(self, index: int):
if not 0 <= index < self._num_examples:
raise IndexError(index)
# Read biopython structure
file_path = self._file_list[index]
structure = fo.read_sdf(str(file_path),
sanitize=False,
add_hs=False,
remove_hs=False)
# assemble the item (no bonds)
item = {
'atoms': fo.bp_to_df(structure),
'id': structure.id,
'file_path': str(file_path),
}
# Add bonds if included
if self._read_bonds:
mol = fo.read_sdf_to_mol(str(file_path),
sanitize=False,
add_hs=False,
remove_hs=False)
bonds_df = fo.get_bonds_list_from_mol(mol[0])
item['bonds'] = bonds_df
if self._transform:
item = self._transform(item)
return item
def __init__(self, csv_file, sdf_file, name='molecules'):
"""Initializes a data set from a CSV file.
Args:
csv_file (str): CSV file with label data.
sdf_file (str): SDF file with coordinates.
name (str, opt.): Name of the dataset. Default: 'molecules'.
"""
df = pd.read_csv(csv_file)
self.raw_data = [df[col] for col in df.keys()]
self.raw_mols = dt.read_sdf_to_mol(sdf_file, sanitize=False)
# Simple sanity check:
# Is the number of molecules the same in both files?
assert len(self.raw_mols) == len(self.raw_data[0])
self.num_atoms = []
self.charges = []
self.positions = []
self.index = []
self.data = []
self.data_keys = [k for k in df.keys()]
for im, m in enumerate(self.raw_mols):
if m is None:
print('Molecule', im + 1, 'could not be processed.')
continue
new_atnums = np.array([a.GetAtomicNum() for a in m.GetAtoms()])
conf_coord = dt.get_coordinates_of_conformer(m)
self.num_atoms.append(m.GetNumAtoms())
self.index.append(im + 1)
self.charges.append(new_atnums)
self.positions.append(conf_coord)
self.data.append([col[im] for col in self.raw_data])
return
def process(self):
label_file = os.path.join(self.root, 'pdbbind_refined_set_labels.csv')
label_df = pd.read_csv(label_file)
i = 0
for raw_path in self.raw_paths:
pdb_code = fi.get_pdb_code(raw_path)
y = torch.FloatTensor([get_label(pdb_code, label_df)])
if '_ligand' in raw_path:
mol_graph = graph.mol_to_graph(
dt.read_sdf_to_mol(raw_path, add_hs=True)[0])
elif '_pocket' in raw_path:
prot_graph = graph.prot_df_to_graph(
dt.bp_to_df(dt.read_any(raw_path, name=pdb_code)))
node_feats, edge_index, edge_feats, pos = graph.combine_graphs(
prot_graph, mol_graph, edges_between=True)
data = Data(node_feats, edge_index, edge_feats, y=y, pos=pos)
data.pdb = pdb_code
torch.save(
data,
os.path.join(self.processed_dir, 'data_{}.pt'.format(i)))
i += 1
else:
continue
def __init__(self, csv_file, sdf_file, name='molecules'):
"""Initializes a data set from a CSV file.
Args:
csv_file (str): CSV file with label data.
sdf_file (str): SDF file with coordinates.
name (str, opt.): Name of the dataset. Default: 'molecules'.
"""
df = pd.read_csv(csv_file)
self.raw_data = [df[col] for col in df.keys()]
self.raw_mols = dt.read_sdf_to_mol(sdf_file, sanitize=False)
# Simple sanity check:
# Is the number of molecules the same in both files?
assert len(self.raw_mols) == len(self.raw_data[0])
self.index = []
self.data = []
self.data_keys = [k for k in df.keys()]
self.mol_dfs = []
for im, m in enumerate(self.raw_mols):
if m is None:
print('Molecule', im + 1, 'could not be processed.')
continue
self.index.append(im + 1)
self.data.append([col[im] for col in self.raw_data])
new_mol_df = dt.mol_to_df(m,
add_hs=False,
structure=df['mol_id'][im])
self.mol_dfs.append(new_mol_df)
return
import numpy as np
import pandas as pd
import atom3d.util.formats as dt
in_dir_name = '../../data/qm9/raw'
csv_file = in_dir_name + '/gdb9.sdf.csv'
sdf_file = in_dir_name + '/gdb9.sdf'
out_file = in_dir_name + '/gdb9_with_cv_atom.csv'
df = pd.read_csv(csv_file)
raw_data = [df[col] for col in df.keys()]
raw_mols = dt.read_sdf_to_mol(sdf_file, sanitize=False)
raw_num_atoms = []
for im, m in enumerate(raw_mols):
if m is None:
print('Molecule', im + 1, 'could not be processed.')
new_numat = None
continue
new_numat = m.GetNumAtoms()
raw_num_atoms.append(new_numat)
raw_num_atoms = np.array(raw_num_atoms)
df['cv_atom'] = df['cv'] - (raw_num_atoms * 2.981)
print('cv:')
print(' mean:', np.mean(df['cv']))
print(' sdev:', np.std(df['cv']))
print('cv_atom:')
print(' mean:', np.mean(df['cv_atom']))
