def test_padmet_to_padmet_reversibility():
"""
Test an issue encountered when a reaction is defined in a direction in one padmet.
And in another the same reaction is defined as reversible.
In old padmet this leads to the reaction having all of its reactants/products as reactants and also as products.
"""
# Read padmet file
padmet_to_padmet('test_data/padmet', 'fabo.padmet')
expected_padmet = PadmetSpec('fabo.padmet')
reactants = [
rlt.id_out
for rlt in expected_padmet.dicOfRelationIn['ENOYL-COA-HYDRAT-RXN']
if rlt.type in ['consumes']
]
products = [
rlt.id_out
for rlt in expected_padmet.dicOfRelationIn['ENOYL-COA-HYDRAT-RXN']
if rlt.type in ['produces']
]
assert sorted(reactants) == [
'TRANS-D2-ENOYL-COA', 'WATER'
] or sorted(products) == ['TRANS-D2-ENOYL-COA', 'WATER']
assert sorted(products) == [
'L-3-HYDROXYACYL-COA'
] or sorted(reactants) == ['L-3-HYDROXYACYL-COA']
os.remove('fabo.padmet')
def test_padmet_to_padmet():
# Using inpu data, create 2 padmets and delete one reaction in each.
fabo_1_padmetSpec = from_pgdb_to_padmet('test_data/pgdb',
extract_gene=True)
fabo_1_padmetSpec.delNode('ACYLCOASYN-RXN')
fabo_1_padmetSpec.generateFile('fabo_1.padmet')
_, _, all_rxns, _ = extract_data_padmet(fabo_1_padmetSpec)
assert not set(FABO_RXNS).issubset(set(all_rxns))
fabo_2_padmetSpec = from_pgdb_to_padmet('test_data/pgdb',
extract_gene=True)
fabo_2_padmetSpec.delNode('ACYLCOADEHYDROG-RXN')
fabo_2_padmetSpec.generateFile('fabo_2.padmet')
_, _, all_rxns, _ = extract_data_padmet(fabo_2_padmetSpec)
assert not set(FABO_RXNS).issubset(set(all_rxns))
# By merging them we should retrieve the two deleted reactions
padmet_to_padmet('fabo_1.padmet,fabo_2.padmet', 'fabo.padmet')
expected_padmet = PadmetSpec('fabo.padmet')
_, _, all_rxns, _ = extract_data_padmet(expected_padmet)
assert set(FABO_RXNS).issubset(set(all_rxns))
os.remove('fabo_1.padmet')
os.remove('fabo_2.padmet')
os.remove('fabo.padmet')
def main():
args = docopt.docopt(__doc__)
if args["--padmetRef"]:
padmetRef = PadmetRef(args["--padmetRef"])
else:
padmetRef = None
output = args["--output"]
verbose = args["-v"]
to_add = args["--to_add"]
padmet_to_padmet.padmet_to_padmet(to_add, output, padmetRef, verbose)
def run_merge(run_id, nb_cpu_to_use, verbose, veryverbose=None):
if verbose:
print('--- Running merge step ---')
merge_start_time = time.time()
aucome_pool = Pool(nb_cpu_to_use)
config_data = parse_config_file(run_id)
padmet_from_annotation_path = config_data['padmet_from_annotation_path']
padmet_from_networks_path = config_data['padmet_from_networks_path']
sbml_from_networks_path = config_data['sbml_from_networks_path']
database_path = config_data['database_path']
structural_padmets_path = config_data['structural_padmets_path']
orthofinder_filtered_path = config_data['orthofinder_filtered_path']
orthofinder_padmet_path = config_data['orthofinder_padmet_path']
padmet_from_annotation_path = config_data['padmet_from_annotation_path']
networks_path = config_data['networks_path']
structural_padmets = [padmet for padmet in os.listdir(structural_padmets_path) if padmet.endswith('.padmet')]
orthofinder_filtered_padmets = [padmet for padmet in os.listdir(orthofinder_filtered_path) if padmet.endswith('.padmet')]
orthofinder_padmets = [padmet for padmet in os.listdir(orthofinder_padmet_path) if padmet.endswith('.padmet')]
pathway_tools_padmets = [padmet for padmet in os.listdir(padmet_from_annotation_path) if padmet.endswith('.padmet')]
if len(structural_padmets) > 0:
padmets = [(padmet, structural_padmets_path + '/' + padmet) for padmet in structural_padmets]
elif len(orthofinder_filtered_padmets) > 0:
padmets = [(padmet, orthofinder_filtered_path + '/' + padmet) for padmet in orthofinder_filtered_padmets]
elif len(orthofinder_padmets) > 0:
padmets = [(padmet, orthofinder_padmet_path + '/' + padmet) for padmet in orthofinder_padmets]
elif len(pathway_tools_padmets) > 0:
padmets = [(padmet, padmet_from_annotation_path + '/' + padmet) for padmet in pathway_tools_padmets]
else:
sys.exit('No padmets have been created, run reconstruction or workflow.')
study_draft_data = []
for study_name, padmet_path in padmets:
tmp_study_data = {'padmet_path': padmet_path, 'study_padmet': study_name, 'padmet_from_networks_path': padmet_from_networks_path,
'sbml_from_networks_path': sbml_from_networks_path, 'database_path': database_path,
'verbose': verbose, 'veryverbose': veryverbose}
study_draft_data.append(tmp_study_data)
aucome_pool.map(create_output, study_draft_data)
aucome_pool.close()
aucome_pool.join()
padmet_to_padmet.padmet_to_padmet(padmet_from_networks_path, networks_path + '/panmetabolism.padmet', verbose=veryverbose)
sbmlGenerator.padmet_to_sbml(padmet=networks_path + '/panmetabolism.padmet', output=networks_path + '/panmetabolism.sbml', verbose=veryverbose)
merge_end_time = (time.time() - merge_start_time)
integer_part, decimal_part = str(merge_end_time).split('.')
merge_time = ".".join([integer_part, decimal_part[:3]])
if verbose:
print("--- merge step done in: %ss ---" %merge_time)
def test_padmet_to_padmet_cli():
# Using inpu data, create 2 padmets and delete one reaction in each.
subprocess.call([
'padmet', 'pgdb_to_padmet', '--pgdb', 'test_data/pgdb', '--output',
'test.padmet', '--extract-gene'
])
fabo_1_padmetSpec = PadmetSpec('test.padmet')
fabo_1_padmetSpec.delNode('ACYLCOASYN-RXN')
fabo_1_padmetSpec.generateFile('fabo_1.padmet')
_, _, all_rxns, _ = extract_data_padmet(fabo_1_padmetSpec)
os.remove('test.padmet')
assert not set(FABO_RXNS).issubset(set(all_rxns))
subprocess.call([
'padmet', 'pgdb_to_padmet', '--pgdb', 'test_data/pgdb', '--output',
'test.padmet', '--extract-gene'
])
fabo_2_padmetSpec = PadmetSpec('test.padmet')
fabo_2_padmetSpec.delNode('ACYLCOADEHYDROG-RXN')
fabo_2_padmetSpec.generateFile('fabo_2.padmet')
_, _, all_rxns, _ = extract_data_padmet(fabo_2_padmetSpec)
os.remove('test.padmet')
assert not set(FABO_RXNS).issubset(set(all_rxns))
# By merging them we should retrieve the two deleted reactions
padmet_to_padmet('fabo_1.padmet,fabo_2.padmet', 'fabo.padmet')
subprocess.call([
'padmet', 'padmet_to_padmet', '--to_add',
'fabo_1.padmet,fabo_2.padmet', '--output', 'fabo.padmet'
])
expected_padmet = PadmetSpec('fabo.padmet')
_, _, all_rxns, _ = extract_data_padmet(expected_padmet)
assert set(FABO_RXNS).issubset(set(all_rxns))
os.remove('fabo_1.padmet')
os.remove('fabo_2.padmet')
os.remove('fabo.padmet')
def analysis_on_group(group_name, groups, config_data, pvclust, nb_cpu_to_use,
verbose):
"""Create reaction dendrogram and extract specific reactions using metabolic networks.
Args:
group_name (str): Name of the group from group_template.tsv.
groups (list): All the species inside the group.
config_data (dict): Dictionary with all configuration paths.
pvclust (boolean): use also pvclust to create reaction dendrogram
nb_cpu_to_use (int): number of CPU for multiprocessing
verbose (bool): Verbose.
"""
database_path = config_data['database_path']
padmetRef = PadmetRef(database_path)
padmet_from_networks_path = config_data['padmet_from_networks_path']
analysis_path = config_data['analysis_path']
all_padmet_path = [
os.path.join(padmet_from_networks_path, name + ".padmet")
for name in groups
]
group_analysis_path = analysis_path + '/' + group_name
if not os.path.isdir(group_analysis_path):
if len(groups) == 1:
sys.exit('A group must contain more than one member.')
for padmet_path in all_padmet_path:
if not os.path.exists(padmet_path):
org_name = os.path.splitext(os.path.basename(padmet_path))[0]
sys.exit(
"Padmet file of organism %s from group %s not found in %s"
% (org_name, group_name, padmet_from_networks_path))
# Compare the padmet to create the reactions.tsv file needed to create the reaction dendrogram.
compare_padmet.compare_padmet(padmet_path=",".join(all_padmet_path),
output=group_analysis_path,
padmetRef=padmetRef,
verbose=verbose,
number_cpu=nb_cpu_to_use)
padmet_to_padmet.padmet_to_padmet(
",".join(all_padmet_path),
group_analysis_path + '/' + group_name + '_panmetabolism.padmet')
sbmlGenerator.padmet_to_sbml(padmet=group_analysis_path + '/' +
group_name + '_panmetabolism.padmet',
output=group_analysis_path + '/' +
group_name + '_panmetabolism.sbml',
verbose=verbose)
dendrogram_reactions_distance.reaction_figure_creation(
reaction_file=group_analysis_path + '/reactions.tsv',
output_folder=group_analysis_path + '/dendrogram_output',
padmetRef_file=database_path,
pvclust=pvclust,
verbose=verbose)
else:
print(
group_analysis_path +
' already exists. Delete it if you want to relaunch the analysis.')
def visu_path_compounds(padmet_pathname,
padmet_ref_pathname,
pathway_ids,
output_file,
hide_currency_metabolites=None):
""" Extract reactions from pathway and create a comppound/reaction graph.
Parameters
----------
padmet_pathname: str
pathname of the padmet file or a folder containing multiple padmet
padmet_ref_pathname: str
pathname of the padmetRef file
pathway_ids: str
name of the pathway (can be multiple pathways separated by a ',')
output_file: str
pathname of the output picture (extension can be .png or .svg)
hide_currency_metabolites: bool
hide currency metabolites
"""
if os.path.isfile(padmet_pathname):
padmet = PadmetSpec(padmet_pathname)
else:
padmet = padmet_to_padmet.padmet_to_padmet(padmet_pathname)
padmet_ref = PadmetRef(padmet_ref_pathname)
pathway_ids = pathway_ids.split(',')
pwy_all_reactions = []
if hide_currency_metabolites:
compounds_to_hide = [
"PROTON", "WATER", "OXYGEN-MOLECULE", "NADP", "NADPH", "ATP",
"PPI", "CARBON-DIOXIDE", "Pi", "ADP", "CO-A", "UDP", "NAD", "NADH",
"AMP", "AMMONIA", "HYDROGEN-PEROXIDE", "Acceptor", "Donor-H2",
"3-5-ADP", "GDP", "CARBON-MONOXIDE", "GTP", "FAD"
]
else:
compounds_to_hide = []
def get_reactions(pathway_id, padmet_ref, pwy_all_reactions):
all_reactions = [
rlt.id_in
for rlt in padmet_ref.dicOfRelationOut.get(pathway_id, None)
if rlt.type == "is_in_pathway"
]
for reaction_id in all_reactions:
if reaction_id in padmet_ref.dicOfNode:
node_reaction = padmet_ref.dicOfNode[reaction_id]
if node_reaction.type == "pathway":
pwy_all_reactions = get_reactions(node_reaction.id,
padmet_ref,
pwy_all_reactions)
else:
if reaction_id not in pwy_all_reactions:
pwy_all_reactions.append(reaction_id)
return pwy_all_reactions
for pathway_id in pathway_ids:
if pathway_id in padmet_ref.dicOfNode:
tmp_pwy_all_reactions = []
tmp_pwy_all_reactions = get_reactions(pathway_id, padmet_ref,
tmp_pwy_all_reactions)
pwy_all_reactions.extend(tmp_pwy_all_reactions)
else:
print("Pathway " + pathway_id + " not in PadmetRef " +
padmet_ref_pathname)
reactions_in_network = []
for reaction_id in pwy_all_reactions:
if reaction_id in padmet.dicOfNode:
reactions_in_network.append(reaction_id)
DG = nx.DiGraph()
custom_node_color = OrderedDict()
for reaction_id in pwy_all_reactions:
# Reaction colors
if reaction_id in reactions_in_network:
custom_node_color[reaction_id] = "lightgreen"
else:
custom_node_color[reaction_id] = "red"
# Reactants & products for each reaction
reactants = [
rlt.id_out
for rlt in padmet_ref.dicOfRelationIn.get(reaction_id, None)
if rlt.type == "consumes"
]
products = [
rlt.id_out
for rlt in padmet_ref.dicOfRelationIn.get(reaction_id, None)
if rlt.type == "produces"
]
for reac in reactants:
if reac not in compounds_to_hide:
if reac not in custom_node_color:
custom_node_color[reac] = "skyblue"
DG.add_edge(reac, reaction_id)
if 'REVERSIBLE' in padmet_ref.dicOfNode[reaction_id].misc[
'DIRECTION']:
DG.add_edge(reaction_id, reac)
for prod in products:
if prod not in compounds_to_hide:
if prod not in custom_node_color:
custom_node_color[prod] = "skyblue"
DG.add_edge(reaction_id, prod)
if 'REVERSIBLE' in padmet_ref.dicOfNode[reaction_id].misc[
'DIRECTION']:
DG.add_edge(prod, reaction_id)
# https://networkx.github.io/documentation/latest/reference/generated/networkx.drawing.nx_pylab.draw_networkx.html
# apt-get install graphviz graphviz-dev (python-pygraphviz)
# pip install pygraphviz
nx.draw_networkx(
DG,
pos=graphviz_layout(DG, prog='neato'), # Layout from graphviz
node_size=1600,
arrows=True,
font_size=11, # font-size for labels
node_shape='s', # shape of nodes
alpha=0.6, # node & edge transparency
width=1.5, # line width for edges
nodelist=list(custom_node_color.keys()),
node_color=[
custom_node_color[node] for node in list(custom_node_color.keys())
])
plt.axis('off')
plt.savefig(output_file, bbox_inches='tight')
plt.clf()
def visu_path_pathways(padmet_pathname, padmet_ref_pathname, pathway_ids,
output_file):
""" Extract reactions from pathway and create a comppound/reaction graph.
Parameters
----------
padmet_pathname: str
pathname of the padmet file or a folder containing multiple padmet
padmet_ref_pathname: str
pathname of the padmetRef file
pathway_ids: str
name of the pathway (can be multiple pathways separated by a ',')
output_file: str
pathname of the output picture (extension can be .png or .svg)
hide_compounds: bool
hide common compounds (like water or proton)
"""
if os.path.isfile(padmet_pathname):
padmet = PadmetSpec(padmet_pathname)
else:
padmet = padmet_to_padmet.padmet_to_padmet(padmet_pathname)
padmet_ref = PadmetRef(padmet_ref_pathname)
# Check if the padmets and padmetref contain the INPUT-COMPOUNDS and OUTPUT-COMPOUNDS in pathway node.misc needed for this analysis.
padmetref_input_compounds_in_pwys = [
1 for node_pathway in padmet_ref.dicOfNode
if padmet_ref.dicOfNode[node_pathway].type == 'pathway'
and 'INPUT-COMPOUNDS' in padmet_ref.dicOfNode[node_pathway].misc
]
padmetref_output_compounds_in_pwys = [
1 for node_pathway in padmet_ref.dicOfNode
if padmet_ref.dicOfNode[node_pathway].type == 'pathway'
and 'OUTPUT-COMPOUNDS' in padmet_ref.dicOfNode[node_pathway].misc
]
if sum(padmetref_input_compounds_in_pwys) == 0 or sum(
padmetref_output_compounds_in_pwys) == 0:
sys.exit(
"The padmetref " + padmet_ref_pathname +
" does not contain INPUT-COMPOUNDS and OUTPUT-COMPOUNDS in the pathway node, can't produce the pathway visualization."
)
padmet_input_compounds_in_pwys = [
1 for node_pathway in padmet.dicOfNode
if padmet.dicOfNode[node_pathway].type == 'pathway'
and 'INPUT-COMPOUNDS' in padmet.dicOfNode[node_pathway].misc
]
padmet_output_compounds_in_pwys = [
1 for node_pathway in padmet.dicOfNode
if padmet.dicOfNode[node_pathway].type == 'pathway'
and 'OUTPUT-COMPOUNDS' in padmet.dicOfNode[node_pathway].misc
]
if sum(padmet_input_compounds_in_pwys) == 0 or sum(
padmet_output_compounds_in_pwys) == 0:
sys.exit(
"The padmet " + padmet_pathname +
" does not contain INPUT-COMPOUNDS and OUTPUT-COMPOUNDS in the pathway node, can't produce the pathway visualization."
)
# Extract pathway from superpathways.
pathway_ids = pathway_ids.split(',')
all_pathways = []
def get_pathways(pathway_id, padmet_ref, pwy_all_reactions):
all_reactions_pathways = [
rlt.id_in
for rlt in padmet_ref.dicOfRelationOut.get(pathway_id, None)
if rlt.type == "is_in_pathway"
]
for reaction_pathway_id in all_reactions_pathways:
if reaction_pathway_id in padmet_ref.dicOfNode:
node_reaction = padmet_ref.dicOfNode[reaction_pathway_id]
if node_reaction.type == "pathway":
pwy_all_reactions.append(reaction_pathway_id)
pwy_all_reactions = get_pathways(node_reaction.id,
padmet_ref,
pwy_all_reactions)
return pwy_all_reactions
for pathway_id in pathway_ids:
if pathway_id in padmet_ref.dicOfNode:
tmp_pwy_all_pathways = []
tmp_pwy_all_pathways = get_pathways(pathway_id, padmet_ref,
tmp_pwy_all_pathways)
all_pathways.extend(tmp_pwy_all_pathways)
else:
print("Pathway " + pathway_id + " not in PadmetRef " +
padmet_ref_pathname)
# Find pathway in the padmet file.
pathways_in_network = []
for pathway_id in all_pathways:
if pathway_id in padmet.dicOfNode:
pathways_in_network.append(pathway_id)
# Create the graph.
DG = nx.DiGraph()
custom_node_color = OrderedDict()
for pwy in all_pathways:
node_pathway = padmet_ref.dicOfNode[pwy]
if pwy in pathways_in_network:
custom_node_color[pwy] = "lightgreen"
else:
custom_node_color[pwy] = "red"
if 'INPUT-COMPOUNDS' in node_pathway.misc and 'OUTPUT-COMPOUNDS' in node_pathway.misc:
for reactant in node_pathway.misc['INPUT-COMPOUNDS'][0].split(','):
if reactant not in custom_node_color:
custom_node_color[reactant] = "skyblue"
DG.add_edge(reactant, pwy)
for product in node_pathway.misc['OUTPUT-COMPOUNDS'][0].split(','):
if product not in custom_node_color:
custom_node_color[product] = "skyblue"
DG.add_edge(pwy, product)
# https://networkx.github.io/documentation/latest/reference/generated/networkx.drawing.nx_pylab.draw_networkx.html
# apt-get install graphviz graphviz-dev (python-pygraphviz)
# pip install pygraphviz
nx.draw_networkx(
DG,
pos=graphviz_layout(DG, prog='neato'), # Layout from graphviz
node_size=1600,
arrows=True,
font_size=11, # font-size for labels
node_shape='s', # shape of nodes
alpha=0.6, # node & edge transparency
width=1.5, # line width for edges
nodelist=list(custom_node_color.keys()),
node_color=[
custom_node_color[node] for node in list(custom_node_color.keys())
])
plt.axis('off')
plt.savefig(output_file, bbox_inches='tight')
plt.clf()