Homec4science
Diffusion Transformation Product Prediction (master) Browse

TP_predict/TP_prediction/find_best_TPs.py
b3fc7cd5bb74master

Last Change
Raw File

find_best_TPs.py
View Options

import sys
sys.path.insert(0, '../src/envipath-python/enviPath_python/')
sys.path.insert(0, '../src/envipath-python/')
from enviPath_python.enviPath import *
from enviPath_python.enviPath import *
from util import *
import getpass
#---------------------------#
# enviPath SETTINGS #
#---------------------------#
# Define the instance to use
INSTANCE_HOST = 'https://envipath.org'
USERNAME = '' # TODO USER: enter your username
# MODEL # TODO USER: Select model for relative reasoning. Default: Standard model BBD - ML - ECC - 2022
# New enviPath - default relative reasoning model on envipath: BBD - ML - ECC - 2022
EP_MODEL_ID = 'https://envipath.org/package/de0cdca1-c3ff-44ed-8ffd-f29c269bfa55/relative-reasoning/646afb6c-6cfc-4d4b-8d22-e196d849ec73'
# New enviPath - BBD+SOIL relative reasoning model on envipath: BBD+SOIL - ML - ECC - 2022
# rr_id = 'https://envipath.org/package/de0cdca1-c3ff-44ed-8ffd-f29c269bfa55/relative-reasoning/0aec0115-941d-4e33-a844-4431d3ec598d'
# New enviPath - BBD+SOIL relative reasoning model on envipath: BBD+SLUDGE - ML - ECC - 2022
# rr_id = 'https://envipath.org/package/de0cdca1-c3ff-44ed-8ffd-f29c269bfa55/relative-reasoning/2a41e599-f962-4268-b7d1-5f1cb252b937'
# New enviPath - BBD+SOIL relative reasoning model on envipath: BBD+SOIL+SLUDGE - ML - ECC - 2022
# rr_id = 'https://envipath.org/package/de0cdca1-c3ff-44ed-8ffd-f29c269bfa55/relative-reasoning/76bd8654-e02f-4fbd-98df-bf61411f9b92'
# PACKAGE - # TODO USER: prepare a new package (manually) and add it's URI here - make sure it is empty when running script!
EP_PACKAGE_ID = 'https://envipath.org/package/de0cdca1-c3ff-44ed-8ffd-f29c269bfa55' # Test package
# List of output packages used for Sludge TP paper
# pkg_id_1 = 'https://envipath.org/package/0915fad3-b889-4aa8-ac98-0707b717be57' # Package for results using BBD - ML - ECC - 2022 model
# pkg_id_2 = 'https://envipath.org/package/80cf58b1-21e2-4c28-9cc6-dc69c6445bdf' # Package for results using BBD+SOIL - ML - ECC - 2022 model
# pkg_id_3 = 'https://envipath.org/package/7d64aa85-2e3c-413f-a538-4d5f2bfd4662' # Package for results using BBD+SLUDGE - ML - ECC - 2022 model
# pkg_id_4 = 'https://envipath.org/package/11f2acd5-5209-4d49-ad77-93f6f6965886' # Package for results using BBD+SOIL+SLUDGE - ML - ECC - 2022 model
#---------------------------#
# PATHWAY SEARCH SETTINGS #
#---------------------------#
# These are the default settings used for the Sludge TP paper.
# They can be modified to direct the pathway search towards a specific objective.
# Maximum number of TPs to predict
MAX_TP = 50
# Lower probability threshold
PROBABILITY_THRESHOLD = -1 # any value equal to or lower than the threshold will be excluded
# Set probabilities of 0 to 0.01 to continue having a weighting scheme downstream of the pathway
INCLUDE_0_PROBABILITIES = False
# Follow moiety - only compounds containing moiety in SMILES will be expanded
MOIETY = "" # e.g., "C(F)(F)F"
# To prioritize small compounds in the queue
SORT_TPS_BY_SIZE = False
# Follow labeled atoms
FOLLOW_LABELED_ATOM = False
ATOM_LABEL = '14'
#---------------------------#
# FILE PATH SETTINGS #
#---------------------------#
# Input/output files
INPUT_FILE_PATH = 'input/input_structures.tsv'
OUTPUT_DIRECTORY = 'output/'
OUTPUT_FILE_TAG = 'TEST'
#---------------------------#
# CONNECT TO ENVIPATH #
#---------------------------#
eP = enviPath(INSTANCE_HOST)
password = getpass.getpass()
eP.login(USERNAME, password)
#---------------------------#
# FUNCTIONS #
#---------------------------#
def __main__(rr_id, pkg_id, tag):
"""
Main function, predicts pathways for a list of input smiles
Output: pathways are saved to specified enviPath package, TP list as .tsv file to output folder
:param rr_id: URI of enviPath relative reasoning mode to be used
:param pkg_id: URI of enviPath package to store resulting pathways
:param tag: string tag to attach to output files for identification
"""
rr = RelativeReasoning(eP.requester, id=rr_id)
pkg = Package(eP.requester, id=pkg_id)
input_list = load_input(INPUT_FILE_PATH)
output_file_path = '{}TP_prediction_{}_top_{}.tsv'.format(OUTPUT_DIRECTORY, tag, MAX_TP)
outfile = open(output_file_path, 'w')
for compound_input in input_list:
result = predict_TPs(compound_input['smiles'], compound_input['name'], rr)
result_list = clean_result(result) # sort and name TPs
pathway_info = upload_envipath_pathway(eP, result_list, pkg)
print_result(outfile, result_list, pathway_info) # continuous writing of result file
def print_result(open_file, result, pathway = None):
"""
Prints output to open file
:param open_file: writable file object
:param result: clean result dictionary from predict_TPs() function
:param pathway: pathway URI from upload_envipath_pathway() function
"""
open_file.write('///') # signifies new pathway entry
if pathway:
open_file.write(' Pathway name: {}, Pathway id: {}'.format(pathway['name'], pathway['id']))
open_file.write('\n')
header= ('SMILES\tname\tcombined_probability\trules\tgeneration\tprobability\tparent\n')
open_file.write(header)
for TP in result:
open_file.write('{}\t{}\t{}\t{}\t{}\t{}\t{}\n'.format(TP['smiles'],
TP['name'],
TP['combined_probability'],
TP['rules'],
TP['generation'],
TP['probability'],
TP['parent'])
)
def predict_TPs(input_smiles, input_name, rr):
"""
Pathway prediction for single compound
:param input_smiles: input smiles of parent compound
:param input_name: name of parent compound
:param rr: relative reasoning object
:return: dictionary of resulting TPs
"""
print('\n### PREDICT TPs FOR COMPOUND {} ###\n'.format(input_name))
num_TP = -1 # counter starts at -1, because source compound is also in the TP list
validated_TPs = {} # container for resulting predictions
queued_items = [{'probability': 1, 'combined_probability': 1, 'smiles': input_smiles, 'generation': 0, 'parent': '',
'rules': '', 'name': input_name, 'size': len(input_smiles)}]
queue = [input_smiles] # queue is updated after each cycle to have top TP first, list of smiles
while num_TP < MAX_TP:
if len(queue) == 0:
print('\nEmpty queue - The exploration of has converged at {} predicted TPs'.format(num_TP))
return validated_TPs # stop TP prediction
smiles = queue.pop(0) # get top item in queue
data = queued_items.pop(0) # remove data from queued items
result_list = expand_smiles(smiles, rr) # create children
TP_dict = result_to_compound_dict(result_list)
queue, queued_items, validated_TPs = update_queue(queue, queued_items, validated_TPs, TP_dict, data)
validated_TPs[smiles] = data
num_TP += 1
return validated_TPs
def update_queue(_queue,_queued_items, _validated_TPs, _TPs, _parent_data):
"""
Update queue with TPs predicted in current iteration
:param _queue: ordered list of smiles to explore
:param _queued_items: ordered list of compound dictionaries, same order as _queue
:param _validated_TPs: list of already validated TPs for resulting pathway
:param _TPs: predicted TPs from current iteration, to be evaluated and added to queue
:param _parent_data: compound dictionary of the parent compound of _TPs
:return: new_queue: new ordered list of smiles to explore
:return _queued_items: new ordered list of compound dictionaries
:return: _validated_TPs: updated list of already validated TPs
"""
parent_probability = _parent_data['combined_probability']
parent_generation = _parent_data['generation']
parent_smiles = _parent_data['smiles']
queue_before = len(_queue)
for smiles in _TPs:
data = _TPs[smiles]
# If the probability is 0 , we don't consider the TP further
this_probability = data['probability']
if this_probability <= PROBABILITY_THRESHOLD:
continue
# If a moiety is given and it is not in SMILES, we don't follow the TP further
if MOIETY not in smiles:
continue
if FOLLOW_LABELED_ATOM and ATOM_LABEL not in smiles:
continue
if INCLUDE_0_PROBABILITIES and this_probability == 0:
this_probability = 0.01
# add combined probability
this_combined_probability = parent_probability * this_probability
this_generation = parent_generation + 1
rules = data['rules']
# first, check if compound already in validated. if yes, update
if smiles in _validated_TPs.keys():
_validated_TPs[smiles] = update_compound_entry(_validated_TPs[smiles],
this_combined_probability, rules,
this_generation, parent_smiles, size_metric='size',
size_value=len(smiles))
# next, check if compound is already in queue. if yes, update
elif smiles in _queue:
index = _queue.index(smiles)
assert smiles == _queued_items[index]['smiles'], \
'smiles {} does not match smiles in {}'.format(smiles, _queued_items[index])
_queued_items[index] = update_compound_entry(_queued_items[index],
this_combined_probability, rules,
this_generation, parent_smiles, size_metric='size',
size_value=len(smiles))
# else, add new item to queue
else:
data['combined_probability'] = this_combined_probability
data['generation'] = this_generation
data['parent'] = parent_smiles
data['carbon_count'] = smiles.upper().count('C')
_queued_items.append(data)
_queue.append(smiles)
assert len(_queued_items) == len(_queue)
# First sort by size
if SORT_TPS_BY_SIZE:
_queued_items.sort(reverse=False, key=lambda x: x['size'])
# order dict by combined probability
_queued_items.sort(reverse=True, key=lambda x: x['combined_probability'])
queue_after = len(_queue)
print('Added {} smiles to queue'.format(queue_after - queue_before))
new_queue = [] # resetting queue
[new_queue.append(x['smiles']) for x in _queued_items]
print ('New queue for compound', parent_smiles)
for q in new_queue:
print(q, _queued_items[new_queue.index(q)]['combined_probability'])
return new_queue, _queued_items, _validated_TPs
#---------------------------#
# MAIN #
#---------------------------#
__main__(rr_id = EP_MODEL_ID, pkg_id = EP_PACKAGE_ID, tag = OUTPUT_FILE_TAG)
c4science · Help