def pml_pharmacophore_reader(path, pyrod_pharmacophore, logger):
""" This functions reads LigandScout pharmacophores and stores them in an internal data structure. """
pharmacophore = []
pml_pharmacophore = et.parse(path)
if pml_pharmacophore.getroot().tag == 'pharmacophore':
pml_pharmacophore = pml_pharmacophore.getroot()
else:
pharmacophore_number = len(pml_pharmacophore.findall('pharmacophore'))
if pharmacophore_number > 0:
if pharmacophore_number == 1:
pml_pharmacophore = pml_pharmacophore.findall(
'pharmacophore')[0]
else:
user_prompt = ''
while user_prompt not in ['yes', 'no']:
user_prompt = input(
'Pharmacophore file contains {} pharmacophores. Only one pharmacophore '
'can be processed at a time. Do you want to continue with the first '
'pharmacophore in the pml file? [yes/no]: '.format(
pharmacophore_number))
if user_prompt == 'no':
sys.exit()
elif user_prompt == 'yes':
pml_pharmacophore = pml_pharmacophore.findall(
'pharmacophore')[0]
else:
update_user('Cannot find any pharmacophore in the pml file.',
logger)
sys.exit()
for feature in pml_pharmacophore:
pharmacophore.append(pml_feature(feature, pyrod_pharmacophore, logger))
if pyrod_pharmacophore:
pharmacophore = merge_pyrod_hydrogen_bonds(pharmacophore, logger)
return pharmacophore
def merge_pyrod_hydrogen_bonds(pharmacophore, logger):
""" This function merges hydrogen bond acceptor and donor features into the pyrod features ha2, hd2 and hda. """
valid_features = [
feature for feature in pharmacophore
if feature[1] not in ['ha2', 'hd2', 'hda', 'ev']
]
exclusion_volumes = [
feature for feature in pharmacophore if feature[1] == 'ev'
]
pyrod_hydrogen_bonds = [
feature for feature in pharmacophore
if feature[1] in ['ha2', 'hd2', 'hda']
]
pyrod_hydrogen_bond_ids = set(
[feature[0] for feature in pyrod_hydrogen_bonds])
for index in pyrod_hydrogen_bond_ids:
feature_pair = [
feature for feature in pyrod_hydrogen_bonds if feature[0] == index
]
pyrod_hydrogen_bond = feature_pair[0]
try:
if len(pyrod_hydrogen_bond[5][0]) == 0:
pyrod_hydrogen_bond[5][0] = feature_pair[1][5][0]
elif len(pyrod_hydrogen_bond[5][1]) == 0:
pyrod_hydrogen_bond[5][1] = feature_pair[1][5][1]
valid_features.append(pyrod_hydrogen_bond)
except IndexError:
update_user(
'The given pharmacophore contains incomplete hydrogen bonding features. Feature type is {} '
'but only one interaction partner was found. Either create a new pharmacophore without '
'splitting ha2, hd2 and hda features or set pyrod pharmacophore to false.'
.format(pyrod_hydrogen_bond[1]), logger)
sys.exit()
valid_features.sort(key=operator.itemgetter(0))
return valid_features + exclusion_volumes
def pharmacophore_reader(path, pyrod_pharmacophore, logger):
""" This function reads pharmacophores and translates them into the internal pharmacophore format.
Format:
------------------------------------------------------------------------
0 1 2 3 4 5 6 7
------------------------------------------------------------------------
0 hi M [0.0,0.0,0.0] 1.5 [] 0.0 150.1
1 pi M [0.0,0.0,0.0] 1.5 [] 0.0 30.1
2 ni M [0.0,0.0,0.0] 1.5 [] 0.0 30.1
3 hd M [0.0,0.0,0.0] 1.5 [[3.0,0.0,0.0]] 1.9499999 30.1
4 ha M [0.0,0.0,0.0] 1.5 [[3.0,0.0,0.0]] 1.9499999 30.1
5 hd2 M [0.0,0.0,0.0] 1.5 [[3.0,0.0,0.0],[0.0,3.0,0.0]] 1.9499999 30.1
6 ha2 M [0.0,0.0,0.0] 1.5 [[3.0,0.0,0.0],[0.0,3.0,0.0]] 1.9499999 30.1
7 hda M [0.0,0.0,0.0] 1.5 [[3.0,0.0,0.0],[0.0,3.0,0.0]] 1.9499999 30.1
8 ai M [0.0,0.0,0.0] 1.5 [[1.0,0.0,0.0]] 0.43633232 30.1
------------------------------------------------------------------------
Legend:
0 - index
1 - type
2 - flag (O - optional, M - mandatory)
3 - core position
4 - core tolerance
5 - partner positions (hda feature with coordinates for first donor than acceptor)
6 - partner tolerance
7 - score
"""
logger.info('Reading pharmacophore from {}.'.format(path))
valid_formats = ['pml', 'pdb']
pharmacophore_format = path.split('.')[-1]
if pharmacophore_format not in valid_formats:
update_user(
'Invalid pharmacophore format detected, only {} and {} are supported.'
.format(', '.join(valid_formats[0:-1]), valid_formats[-1]), logger)
sys.exit()
pharmacophore = []
if pharmacophore_format == 'pml':
pharmacophore = pml_pharmacophore_reader(path, pyrod_pharmacophore,
logger)
elif pharmacophore_format == 'pdb':
pharmacophore = pdb_pharmacophore_reader(path, pyrod_pharmacophore,
logger)
if not pyrod_pharmacophore:
pharmacophore = renumber_features(pharmacophore)
return pharmacophore
def pdb_pharmacophore_reader(path, pyrod_pharmacophore, logger):
""" This functions reads pdb pharmacophores in pyrod format and stores them in an internal data structure. """
if not pyrod_pharmacophore:
update_user(
'This format is specific to pyrod pharmacophores. Make sure hte format is correct.',
logger)
pharmacophore = []
with open(path, 'r') as pharmacophore_file:
for line in pharmacophore_file.readlines():
if line[:6].strip() == 'ATOM':
if line[12:16].strip() == 'C':
feature_list = [0, 0, 0, 0, 0.0, [], 0.0, 0.0]
feature_list[0] = int(line[22:26])
feature_list[1] = line[17:20].strip()
feature_list[2] = line[21:22].strip()
feature_list[3] = [
float(line[30:38]),
float(line[38:46]),
float(line[46:54])
]
feature_list[4] = float(line[54:60])
feature_list[7] = float(line[60:66])
pharmacophore.append(feature_list)
if 'P' in line[12:16]:
feature_list = [
feature for feature in pharmacophore
if feature[0] == int(line[22:26])
]
pharmacophore = [
feature for feature in pharmacophore
if feature[0] == int(line[22:26])
]
feature_list[5].append([
float(line[30:38]),
float(line[38:46]),
float(line[46:54])
])
feature_list[6] = float(line[54:60])
pharmacophore.append(feature_list)
return pharmacophore
def pml_feature(feature, pyrod_pharmacophore, logger):
""" This function converts a pml feature into the internal pharmacophore feature format. """
translate_features = {
'H': 'hi',
'PI': 'pi',
'NI': 'ni',
'HBA': 'ha',
'HBD': 'hd',
'AR': 'ai',
'exclusion': 'ev'
}
feature_list = [0, 0, 0, 0, 0.0, [], 0.0, 0.0]
feature_list[0] = feature.attrib['featureId']
if feature.tag == 'volume':
feature_list[1] = translate_features[feature.attrib['type']]
else:
feature_list[1] = translate_features[feature.attrib['name']]
if pyrod_pharmacophore:
try:
feature_list[0] = int(feature.attrib['featureId'].split('_')[1])
feature_list[1] = feature.attrib['featureId'].split('_')[0]
except (IndexError, ValueError):
update_user(
'You attempted to read a pml pharmacophore, that was not generated by PyRod or with an older '
'version of PyRod. Please set pyrod pharmacophore to false!',
logger)
sys.exit()
feature_list[2] = 'M'
if feature.attrib['optional'] == 'true':
feature_list[2] = 'O'
if feature.tag in ['point', 'volume']:
pml_feature_point_or_volume(feature_list, feature)
elif feature.tag == 'vector':
pml_feature_vector(feature_list, feature)
elif feature.tag == 'plane':
pml_feature_plane(feature_list, feature)
return feature_list
def generate_exclusion_volumes(dmif, directory, debugging, shape_cutoff,
restrictive):
""" This function generates exclusion volumes. The exclusion volumes are described with a list of properties as
follows.
Format:
------------------------------------------------------------------------
0 1 2 3 4 5 6 7
------------------------------------------------------------------------
0 ev M [0.0,0.0,0.0] 1.0 [] 0.0 1.0
1 ev M [2.0,0.0,0.0] 1.0 [] 0.0 1.0
------------------------------------------------------------------------
Legend:
0 - index
1 - type
2 - flag (O - optional, M - mandatory)
3 - core position
4 - core tolerance [not needed for exclusion volumes]
5 - partner positions [not needed for exclusion volumes]
6 - partner tolerance [not needed for exclusion volumes]
7 - weight
"""
logger = setup_logger('exclusion_volumes', directory, debugging)
update_user('Generating exclusion volumes.', logger)
grid_space = 0.5
exclusion_volume_space = 4
if restrictive:
exclusion_volume_space = 2
grid_tree = cKDTree([[x, y, z]
for x, y, z in zip(dmif['x'], dmif['y'], dmif['z'])])
dtype = [('x', float), ('y', float), ('z', float), ('shape', int),
('count', int)]
dmif_shape = np.array(
[(x, y, z, shape, 0)
for x, y, z, shape in zip(dmif['x'], dmif['y'], dmif['z'],
dmif['shape']) if shape < shape_cutoff],
dtype=dtype)
positions = np.array([[
x, y, z
] for x, y, z in zip(dmif_shape['x'], dmif_shape['y'], dmif_shape['z'])])
shape_tree = cKDTree(positions)
shape_grid_size = len(dmif_shape)
# store number of neighbors with shape score smaller than shape_cutoff for grid points
for index in range(shape_grid_size):
dmif_shape['count'][index] = len(
shape_tree.query_ball_point(positions[index], grid_space * 4))
# sort for neighbor count
dmif_shape = np.sort(dmif_shape, order='count')
# rebuild positions and shape_tree
positions = np.array([[
x, y, z
] for x, y, z in zip(dmif_shape['x'], dmif_shape['y'], dmif_shape['z'])])
shape_tree = cKDTree(positions)
used = []
exclusion_volumes = []
counter = 1
start = time.time()
for index in range(shape_grid_size):
# grid_point index should not be in used list
if index not in used:
neighbor_list = shape_tree.query_ball_point(
positions[index], exclusion_volume_space / 2)
# elements of neighbor_list should not be in used list
if len(set(neighbor_list +
used)) == len(neighbor_list) + len(used):
# grid_point should not be at border of grid
if len(
grid_tree.query_ball_point(positions[index],
r=grid_space * 2)) == 33:
# grid_point should not be directly at border of binding pocket
if len(
shape_tree.query_ball_point(positions[index],
r=grid_space)) == 7:
# grid_point should not be surrounded by grid_points outside the binding pocket
if len(
shape_tree.query_ball_point(
positions[index], r=grid_space * 2)) < 33:
exclusion_volumes.append([
counter, 'ev', 'M', positions[index], 1.0, [],
0.0, 1.0
])
counter += 1
used += neighbor_list
eta = ((time.time() - start) / (index + 1)) * (shape_grid_size -
(index + 1))
update_progress(
float(index + 1) / shape_grid_size,
'Progress of exclusion volume generation', eta)
logger.debug('Passed grid index {}.'.format(index))
update_user(
'Finished with generation of {} exclusion volumes.'.format(
len(exclusion_volumes)), logger)
return exclusion_volumes
def generate_library(pharmacophore_path, output_format, library_dict,
library_path, pyrod_pharmacophore, directory, debugging):
""" This function writes a combinatorial pharmacophore library. """
logger = setup_logger('library', directory, debugging)
update_user('Starting library generation.', logger)
template_pharmacophore = pharmacophore_reader(pharmacophore_path,
pyrod_pharmacophore, logger)
pharmacophore_library = []
essential_hb, essential_hi, essential_ai, essential_ii = [], [], [], []
optional_hb, optional_hi, optional_ai, optional_ii = [], [], [], []
exclusion_volumes = []
# analyzing pharmacophore
for index, feature in enumerate(template_pharmacophore):
if feature[1] == 'ev':
exclusion_volumes.append(feature)
else:
if feature[1] in ['ha', 'hd', 'ha2', 'hd2', 'hda']:
if feature[2] == 'O':
optional_hb.append(index)
else:
essential_hb.append(index)
elif feature[1] == 'hi':
if feature[2] == 'O':
optional_hi.append(index)
else:
essential_hi.append(index)
elif feature[1] in ['pi', 'ni']:
if feature[2] == 'O':
optional_ii.append(index)
else:
essential_ii.append(index)
elif feature[1] == 'ai':
if feature[2] == 'O':
optional_ai.append(index)
else:
essential_ai.append(index)
essential_features = essential_hb + essential_hi + essential_ai + essential_ii
for hb_combination in combine_features(
optional_hb,
library_dict['minimal hydrogen bonds'] - len(essential_hb),
library_dict['maximal hydrogen bonds'] - len(essential_hb) + 1):
for hi_combination in combine_features(
optional_hi, library_dict['minimal hydrophobic interactions'] -
len(essential_hi),
library_dict['maximal hydrophobic interactions'] -
len(essential_hi) + 1):
for ai_combination in combine_features(
optional_ai,
library_dict['minimal aromatic interactions'] -
len(essential_ai),
library_dict['maximal aromatic interactions'] -
len(essential_ai) + 1):
for ii_combination in combine_features(
optional_ii,
library_dict['minimal ionizable interactions'] -
len(essential_ii),
library_dict['maximal ionizable interactions'] -
len(essential_ii) + 1):
pharmacophore = (essential_features + hb_combination +
hi_combination + ai_combination +
ii_combination)
if evaluate_pharmacophore(pharmacophore,
template_pharmacophore,
library_dict,
pyrod_pharmacophore):
pharmacophore_library.append(pharmacophore)
# estimate maximal library size and ask user if number and space of pharmacophores is okay
pharmacophore_writer(template_pharmacophore, [output_format],
'template_pharmacophore', library_path, logger)
pharmacophore_library_size = bytes_to_text(
os.path.getsize('{}/{}.{}'.format(
library_path, 'template_pharmacophore', output_format)) *
len(pharmacophore_library))
user_prompt = ''
while user_prompt not in ['yes', 'no']:
user_prompt = input(
'{} pharmacophores will be written taking about {} of space.\n'
'Do you want to continue? [yes/no]: '.format(
len(pharmacophore_library), pharmacophore_library_size))
if user_prompt == 'no':
sys.exit()
start = time.time()
# write pharmacophores
maximal_exclusion_volume_id = max(
[exclusion_volume[0] for exclusion_volume in exclusion_volumes])
for counter, index_pharmacophore in enumerate(pharmacophore_library):
extra_exclusion_volumes = []
extra_ev_counter = 1
pharmacophore = []
for index_feature in index_pharmacophore:
feature = template_pharmacophore[index_feature]
feature[2] = 'M'
pharmacophore.append(feature)
if feature[1] in ['ha', 'hd', 'ha2', 'hd2', 'hda']:
extra_exclusion_volumes.append([
maximal_exclusion_volume_id + extra_ev_counter, 'ev', 'M',
feature[5][0], 1.0, [], 0.0, 0.0
])
extra_ev_counter += 1
if feature[1] in ['ha2', 'hd2', 'hda']:
extra_exclusion_volumes.append([
maximal_exclusion_volume_id + extra_ev_counter, 'ev',
'M', feature[5][1], 1.0, [], 0.0, 0.0
])
extra_ev_counter += 1
pharmacophore_writer(
pharmacophore + exclusion_volumes + extra_exclusion_volumes,
[output_format], str(counter), library_path, logger)
update_progress(
(counter + 1) / len(pharmacophore_library),
'Writing {} pharmacophores'.format(len(pharmacophore_library)),
((time.time() - start) /
(counter + 1)) * (len(pharmacophore_library) - (counter + 1)))
update_user('Wrote pharmacophores to {}.'.format(library_path), logger)
return
description='\n'.join(logo),
formatter_class=argparse.RawTextHelpFormatter)
parser.add_argument('conf', help='path to configuration file')
parser.add_argument('--verbose',
dest='debugging',
action='store_true',
help='verbose logging for debugging')
conf = parser.parse_args().conf
debugging = parser.parse_args().debugging
config = configparser.ConfigParser()
config.read(conf)
directory = config.get('directory', 'directory')
if len(directory) == 0:
directory = os.getcwd() + '/pyrod'
logger = setup_logger('main', directory, debugging)
update_user('\n'.join(logo), logger)
logger.debug('\n'.join(
[': '.join(list(_)) for _ in config.items('directory')]))
# defining grid
if config.has_section('test grid parameters'):
logger.debug('\n'.join([
': '.join(list(_)) for _ in config.items('test grid parameters')
]))
center, edge_lengths, name = test_grid_parameters(config)
# determine space resulting in less than 100000 grid points
space = 0.5
space_found = False
while not space_found:
grid = generate_grid(center, edge_lengths, space)
if len(grid) < 100000:
space_found = True
def ensemble_to_centroid(topology, trajectories, output_name, directory,
debugging):
logger = setup_logger('ensembles_to_centroid', directory, debugging)
output_directory = '/'.join([directory, output_name])
protein_topology = '/'.join([output_directory, 'protein.pdb'])
protein_trajectories = []
frames = []
# check if frames in trajectory files, delete empty trajectory files, collect frames in list
for x in range(len(trajectories)):
with open('{}/frames_{}.csv'.format(output_directory, x), 'r') as csv:
frames += csv.readlines()
os.remove('{}/frames_{}.csv'.format(output_directory, x))
try:
mda.Universe(
protein_topology,
'/'.join([output_directory, 'ensemble_{}.dcd'.format(x)]))
protein_trajectories.append(x)
except OSError:
os.remove('/'.join([output_directory,
'ensemble_{}.dcd'.format(x)]))
if len(protein_trajectories) > 0:
# info to user
if len(frames) > 1:
update_user(
'Getting centroid from {} protein conformations.'.format(
len(frames)), logger)
else:
update_user('Found only 1 protein conformations.', logger)
# merge trajectories into one file
protein_trajectories = [
'/'.join([output_directory, 'ensemble_{}.dcd'.format(x)])
for x in protein_trajectories
]
u = mda.Universe(protein_topology, protein_trajectories)
with mda.Writer('/'.join([output_directory, 'ensemble.dcd']),
n_atoms=u.atoms.n_atoms) as DCD:
for _ in u.trajectory:
DCD.write(u.atoms)
# remove sub trajectories
for path in protein_trajectories:
os.remove(path)
# find centroid of frames
if len(frames) > 1:
u = mda.Universe(protein_topology,
'/'.join([output_directory, 'ensemble.dcd']))
conf_dist_matrix = encore.confdistmatrix.get_distance_matrix(
u,
selection='all',
superimpose=True,
n_job=1,
weights='mass',
metadata=False,
verbose=False)
centroid = conf_dist_matrix.as_array().sum(axis=1).argmin()
else:
centroid = 0
# write centroid
u = mda.Universe(topology,
trajectories[int(frames[centroid].split()[0])])
file_path('centroid.pdb', output_directory)
with mda.Writer('/'.join([output_directory, 'centroid.pdb']),
bonds=None,
n_atoms=u.atoms.n_atoms) as PDB:
for _ in u.trajectory[int(frames[centroid].split()[1]
):int(frames[centroid].split()[1]) + 1:]:
PDB.write(u.atoms)
# write csv with frame references
file_path('frames.csv', output_directory)
frames[centroid] = '{}\t{}\t{}\n'.format(frames[centroid].split()[0],
frames[centroid].split()[1],
'centroid')
with open('{}/frames.csv'.format(output_directory), 'w') as csv:
csv.write(''.join(['trajectory\tframe\tcentroid\n'] + frames))
else:
update_user('No protein conformations found.', logger)
sys.exit()
return
