def _ETKDG(ifile) -> (bool, str):
""" Assigns 3D structures to the molecular structures provided as input.
"""
success_list = [True for i in range(sdfu.count_mols(ifile))]
LOG.info('Converting to ETKDG 3D structures')
try:
suppl = Chem.SDMolSupplier(ifile)
except Exception as e:
LOG.critical('Unable to create supplier')
raise e
# not true, UNABLE TO CREATE SUPPLIER
# return False, 'unable to compute 3D structures'
filename, fileext = os.path.splitext(ifile)
ofile = filename + '_3d' + fileext
LOG.debug(f'3D stucture ouput file is: {ofile}')
with open(ofile, 'w') as fo:
mcount = 0
for mol in suppl:
if mol is None:
LOG.debug('Supplier failed to read'
f' molecule #{mcount+1} in {ifile}')
continue
try:
mol3 = Chem.AddHs(mol)
AllChem.EmbedMolecule(mol3, AllChem.ETKDG())
except:
LOG.error('Failed to generate 3D structures using'
f'ETKDG method for molecule #{mcount+1} in {ifile}')
success_list[mcount]=False
mcount += 1
continue
## debug for testing error handling. This code simulates a 3D conversion error
# if mcount == 3 :
# print ('@ETKDG debug, skipping mol 3')
# LOG.error('Failed to generate 3D structures using'
# f'ETKDG method for molecule #{mcount+1} in {ifile}')
# success_list[mcount]=False
# mcount += 1
# continue
fo.write(Chem.MolToMolBlock(mol3))
fo.write('\n$$$$\n') # end of mol
mcount += 1
return success_list, ofile
def convert3D(self, ifile, method):
'''
Assigns 3D structures to the molecular structures provided as input.
'''
success_list = [True for i in range(sdfu.count_mols(ifile))]
if not method:
return success_list, ifile
if 'ETKDG' in method:
success_list, ofile = convert3D._ETKDG(ifile)
return success_list, ofile
def ionize(self, ifile, method):
'''
Adjust the ionization status of the molecular structure,
using a given pH.
'''
success_list = [True for i in range(sdfutils.count_mols(ifile))]
if not method:
return success_list, ifile
else:
LOG.debug('ionize called, but no method implemented so far')
# methods here
return success_list, ifile
def workflow_series(self, input_file):
'''
Executes in sequence methods required to generate MD,
starting from a single molecular file
input : ifile, a molecular file in SDFile format
output: results contains the following lists
results[0] a numpy bidimensional array containing MD
results[1] a list of strings containing the names of the MD vars
results[2] a list of booleans indicating for which objects the
MD computations succeeded
'''
mol_index = [True for i in range(sdfu.count_mols(input_file))]
###
# 1. normalize
###
success_list, output_normalize_file = self.normalize(
input_file, self.parameters['normalize_method'])
success, mol_index = self.updateMolIndex(mol_index, success_list)
if not success:
return False, 'failed to normalize ' + input_file
###
# 2. ionize
###
success_list, output_ionize_file = self.ionize(
output_normalize_file, self.parameters['ionize_method'])
success, mol_index = self.updateMolIndex(mol_index, success_list)
if not success:
return False, 'failed to ionize ' + input_file
###
# 3. convert3D
###
success_list, output_convert3D_file = self.convert3D(
output_ionize_file, self.parameters['convert3D_method'])
success, mol_index = self.updateMolIndex(mol_index, success_list)
if not success:
return False, 'failed to convert 3D ' + input_file
###
# 4. compute MD
###
success, results = self.computeMD(output_convert3D_file,
self.parameters['computeMD_method'])
if not success:
return False, results
x = results[0]
xnames = results[1]
success_list = results[2]
success, mol_index = self.updateMolIndex(mol_index, success_list)
return success, (x, xnames, mol_index)
def workflow_objects(self, input_file):
'''
Executes in sequence methods required to generate MD,
starting from a single molecular file.
input : ifile, a molecular file in SDFile format
output: results is a numpy bidimensional array containing MD
'''
success_list = []
md_results = []
va_results = []
# split in single molecule pieces
num_mol = sdfu.count_mols(input_file)
success, results = sdfu.split_SDFile(input_file, num_mol)
if not success:
return success, results
file_list = results[0]
file_size = results[1]
# check if any of the molecules is empty
for fsize in file_size:
success_list.append(fsize == 1)
first_mol = True
for i, ifile in enumerate(file_list):
if not success_list[i]: # molecule was empty, do not process
LOG.error(f'Molecule {i+1} in {ifile} is empty, skiping...')
continue
success, results = self.workflow_series(ifile)
# since the workflow was run for a single molecule, results[2] is ignored, because it must match
# the value in success
success_list[i] = success
if not success: # failed in the workflow
LOG.error(f'Workflow failed for molecule #{str(i+1)}'
f' in file {input_file}')
continue
if first_mol: # first molecule
md_results = results[0]
va_results = results[1]
num_var = len(md_results)
first_mol = False
else:
if len(results[0]) != num_var:
LOG.warning(f'MD length for molecule #{str(i+1)} in file'
f' {input_file} does not match the MD length'
'of the first molecule')
success_list[i] = False
continue
md_results = np.vstack((md_results, results[0]))
#print (success_list)
return True, (md_results, va_results, success_list)
def normalize(self, ifile, method):
'''
Generates a simplified SDFile with MolBlock and an internal ID for
further processing
Note that this method is applied to every molecule and that it removes
mol blocks in the input SDFile not able to generate a valid mol
Also, when defined in control, applies chemical standardization
protocols, like the one provided by Francis Atkinson (EBI),
accessible from:
https://github.com/flatkinson/standardiser
Returns a tuple containing the result of the method and (if True)
the name of the output molecule and an error message otherwyse
'''
success_list = [True for i in range(sdfu.count_mols(ifile))]
if not method:
method = ''
LOG.info('Starting normalization...')
try:
suppl = Chem.SDMolSupplier(ifile)
LOG.debug(f'mol supplier created from {ifile}')
except Exception as e:
LOG.error('Unable to create mol supplier with the exception: '
f'{e}')
return False, 'Error at processing input file for standardizing structures'
filename, fileext = os.path.splitext(ifile)
ofile = filename + '_std' + fileext
LOG.debug(f'writing standarized molecules to {ofile}')
with open(ofile, 'w') as fo:
mcount = 0
# merror = 0
for m in suppl:
# molecule not recognised by RDKit
if m is None:
LOG.error('Unable to process molecule'
f' #{mcount+1} in {ifile}')
continue
name = sdfu.getName(m,
count=mcount,
field=self.parameters['SDFile_name'],
suppl=suppl)
parent = None
if 'standardize' in method:
try:
parent = standardise.run(Chem.MolToMolBlock(m))
except standardise.StandardiseException as e:
if e.name == "no_non_salt":
# very commong warning, use parent mol and proceed
LOG.debug(
f'"No non salt error" found. Skiped standardize for mol'
f' #{mcount} {name}')
parent = Chem.MolToMolBlock(m)
else:
# serious issue, no parent was generated, use original mol
if (parent is None):
LOG.error(
f'Critical standardize exception: {e}'
f' when processing mol #{mcount} {name}. Skipping normalization'
)
parent = Chem.MolToMolBlock(m)
# minor isse, parent was generated, show a warning and proceed
else:
LOG.info(
f'Standardize exception: {e}'
f' when processing mol #{mcount} {name}. Normalization applied'
)
#return False, e.name
except Exception as e:
# this error means an execution error running standardizer
# the molecule is discarded and therefore the list of molecules must be updated
LOG.error(
f'Critical standardize execution exception {e}'
f' when processing mol #{mcount} {name}. Discarding molecule'
)
success_list[mcount] = False
continue
else:
LOG.info(f'Skipping normalization.')
parent = Chem.MolToMolBlock(m)
# in any case, write parent plus internal ID (flameID)
fo.write(parent)
# *** discarded method to control errors ****
# flameID = 'fl%0.10d' % mcount
# fo.write('> <flameID>\n'+flameID+'\n\n')
mcount += 1
# terminator
fo.write('$$$$\n')
return success_list, ofile