def print_initial_conditions(cell_lines, gene_names, fname):
"""Print a table with initial states/conditions for each cell line."""
fh = open(fname, 'wt')
genes_per_model = {}
# First we collect the state and initial value of proteins
# in each cell line into a dict
for cell_line in cell_lines:
genes_per_model[cell_line] = {}
model = pklload('pysb_model_%s' % cell_line)
inits = sorted(model.initial_conditions,
key=lambda x: x[0].monomer_patterns[0].monomer.name)
# A list of "basic" states, deviation from which indicates
# a "special" state like a mutated form
basic_states = ['n', 'u', 'WT', 'inactive', None]
# Iterate over all the initial conditions to collect them
# in the genes_per_model datastructure
for cplx, initial in inits:
name = cplx.monomer_patterns[0].monomer.name
extra_states = []
for site, state in cplx.monomer_patterns[0].site_conditions.items(
):
if site == 'loc':
continue
if state not in basic_states:
extra_states.append('%s=%s' % (site, state))
val_str = '%d' % initial.value
if extra_states:
val_str += ' (%s)' % ', '.join(extra_states)
genes_per_model[cell_line][name] = val_str
# Print the file while skipping proteins that aren't in any of
# the cell line specific models.
header = 'Protein\t' + '\t'.join(cell_lines) + '\n'
fh.write(header)
for gene in gene_names:
line_vals = [gene]
no_val = True
for cell_line in cell_lines:
val = genes_per_model[cell_line].get(gene)
if not val:
val = '-'
else:
no_val = False
line_vals.append(val)
line = '\t'.join(line_vals)
if not no_val:
fh.write(line + '\n')
fh.close()
data = read_rppa_data()
antibody_agents = get_antibody_agents()
agents_to_observe = []
for agents in antibody_agents.values():
agents_to_observe += agents
# Get all the data for Task 5
stmts_to_check = get_task_5(data, INVERSE)
dose = 1.0
scored_paths = {}
models = {}
global_mc = None
# Run model checking per cell line
for cell_line in stmts_to_check.keys():
print('Cell line: %s\n=============' % cell_line)
model = pklload('pysb_model_%s' % cell_line)
scored_paths[cell_line] = {}
# Make a Model Checker with the
# - model contextualized to the cell line
# - the statements for the given condition
# - agents for which observables need to be made
global_mc = get_global_mc(model, stmts_to_check, agents_to_observe)
for drug in stmts_to_check[cell_line].keys():
print('Drug: %s\n=============' % drug)
# Get all the statements and values for the condition
stmts_condition, values_condition = \
stmts_to_check[cell_line][drug][dose]
path_results = []
for stmt in stmts_condition:
pr = global_mc.check_statement(stmt, 100, 7)
path_results.append(pr)
import pickle
from indra.sources import trips
from indra.explanation.model_checker import PysbModelChecker
from util import pklload
pysb_model = pklload('pysb_model')
pysb_stmts = pklload('pysb_stmts')
#stmts_to_check = trips.process_text('BRAF phosphorylates MAPK1.').statements
stmts_to_check = trips.process_text('HRAS phosphorylates MAPK1.').statements
mc = PysbModelChecker(pysb_model, stmts_to_check)
#paths = mc.check_model(max_paths=100, max_path_length=6)
from util import pklload
from collections import defaultdict
import indra.tools.assemble_corpus as ac
if __name__ == '__main__':
# Load cached Statements just before going into the model
stmts = pklload('pysb_stmts')
# Start a dictionary for source counts
sources_count = defaultdict(int)
# Count statements according to sources of evidence
for stmt in stmts:
sources = tuple(
sorted(list(set([ev.source_api for ev in stmt.evidence]))))
sources_count[sources] += 1
# Statements from databases only
db_only = 0
# Statements from reading only
reading_only = 0
# Statements from databases and reading
mixture = 0
# Database sources
dbs = set(['bel', 'biopax', 'phosphosite', 'signor'])
# Reader sources
readers = set(['reach', 'trips', 'sparser', 'r3'])
for k, v in sources_count.items():
d = set(k).intersection(dbs)
r = set(k).intersection(readers)
if d and r:
mixture += v
