def extract_data_sbml(sbml_filepath):
reader = libsbml.SBMLReader()
document = reader.readSBML(sbml_filepath)
model = document.getModel()
compounds = model.getListOfSpecies()
reactions = model.getListOfReactions()
genes = []
for reactionSBML in reactions:
notes = sbmlPlugin.parseNotes(reactionSBML)
if "GENE_ASSOCIATION" in list(notes.keys()):
# Using sbmlPlugin to recover all genes associated to the reaction
for gene in sbmlPlugin.parseGeneAssoc(
notes["GENE_ASSOCIATION"][0]):
if gene not in genes:
genes.append(gene)
id_compounds = [
sbmlPlugin.convert_from_coded_id(compound.id)[0]
for compound in compounds
]
id_reactions = [
sbmlPlugin.convert_from_coded_id(reaction.id)[0]
for reaction in reactions
]
return genes, id_compounds, id_reactions
def check_ids(model_metabolic, model_faa, cutoff, verbose=False):
"""
check if genes ids of model_metabolic = model_faa for a given cutoff
faa genes ids are in the first line of each sequence: >GENE_ID ....
metabolic netowkrs genes ids are in note section, GENE_ASSOCIATION: gene_id-1 or gene_id-2
Parameters
----------
model_metabolic: str
path to sbml file
model_faa: str
path to fasta faa file
cutoff: int
cutoff genes ids from model found in faa
verbose: bool
verbose
Returns
-------
bool
True if same ids, if verbose, print % of genes under cutoff
"""
reader = libsbml.SBMLReader()
document = reader.readSBML(model_metabolic)
model = document.getModel()
document.getNumErrors()
listOfReactions = model.getListOfReactions()
#convert to set
model_metabolic_ids = set(itertools.chain.from_iterable([sp.parseGeneAssoc(geneAssoc)
for geneAssoc in (sp.parseNotes(r).get("GENE_ASSOCIATION",[None])[0] for r in listOfReactions)
if geneAssoc is not None]))
with open(model_faa, "r") as f:
model_faa_ids = set([record.id for record in SeqIO.parse(f, "fasta")])
diff_genes = model_metabolic_ids.difference(model_faa_ids)
try:
diff_genes_ratio = float(len(diff_genes))/float(len(model_metabolic_ids))
except ZeroDivisionError:
raise SystemExit("No genes found in model metabolic")
#if all model_metabolic_ids are in model_faa_ids
if diff_genes_ratio == 0:
if verbose: print("all genes of the model_metabolic are in the model_faa")
return True
#if not check if the nb is sup-equal to the cutoff
elif diff_genes_ratio <= float(1-cutoff):
if verbose: print("Only %.2f%% genes of the model_metabolic are not in the model_faa" % (diff_genes_ratio*100))
return True
else:
if verbose:
print("%s%% genes of the model_metabolic are not in the model_faa" % (diff_genes_ratio*100))
print(";".join(diff_genes))
return False
def create_sbml_stat(species_name, sbml_file):
"""Extract reactions/pathways/compounds/genes from a sbml file.
Args:
species_name (str): species names
sbml_file (str): path to a sbml file
Returns
list: [species name, list of genes, list of reactions, list of reactions associated with genes, list of compounds]
"""
tree = etree.parse(sbml_file)
sbml = tree.getroot()
genes = []
reactions = []
gene_associated_rxns = []
compounds = []
for e in sbml:
if e.tag[0] == "{":
uri, tag = e.tag[1:].split("}")
else:
tag = e.tag
if tag == "model":
model_element = e
for els in model_element:
if 'listOfSpecies' in els.tag:
for el in els:
compounds.append(
sbmlPlugin.convert_from_coded_id(el.get('metaid'))[0])
if 'listOfReactions' in els.tag:
for el in els:
reaction_id = sbmlPlugin.convert_from_coded_id(el.get('id'))[0]
reactions.append(reaction_id)
for subel in el.getchildren():
if 'notes' in subel.tag:
for subsubel in subel.getchildren():
for subsubsubel in subsubel.getchildren():
if 'GENE_ASSOCIATION' in subsubsubel.text:
for gene in sbmlPlugin.parseGeneAssoc(
subsubsubel.text):
genes.append(
gene.replace(
'GENE_ASSOCIATION:', ''))
if reaction_id not in gene_associated_rxns:
gene_associated_rxns.append(
reaction_id)
return [species_name, genes, reactions, gene_associated_rxns, compounds]
def add_delete_rxn(data_file, padmetSpec, output, padmetRef=None, source=None, tool=None, category="MANUAL", verbose=False):
"""
Read a data_file (form created with template_add_delete and filed), for each reaction
if column 'Action' == 'add':
add the reaction from padmetRef to padmetSpec.
elif column 'Action' == 'delete':
remove the reaction
Can't add a reaction without a padmetRef !
the source ensure the traceability of the reaction, its a simple tag ex 'pathway_XX_update'
if not given the filename of data_file will be used.
if a tool was used to infer the reaction, define tool='name_of_the_tool'
Parameters
----------
data_file: str
path to file based on template_new_rxn()
padmetSpec: padmet.classes.PadmetSpec
padmet to update
padmetRef: padmet.classes.PadmetRef
padmet containing the database of reference
output: str
path to the new padmet file
source: str
tag associated to the new reactions to create and add, used for traceability
tool: str
The eventual tool used to infer the reactions to create and add
category: str
The default category of the reaction added manually is 'MANUAL'. Must not be changed.
verbose: bool
if True print information
"""
if not source:
filename = os.path.splitext(os.path.basename(data_file))[0]
source = filename
source = source.upper()
if tool:
tool = tool.upper()
if not category:
category = "MANUAL"
with open(data_file, 'r') as csvfile:
dialect = csv.Sniffer().sniff(csvfile.read())
csvfile.seek(0)
reader = csv.reader(csvfile, dialect)
file_name = os.path.basename(data_file)
file_name = os.path.splitext(file_name)[0]
reader = csv.DictReader(csvfile, delimiter=dialect.delimiter)
for row in reader:
element_id, comment, action, genes_assoc = row["idRef"], row["Comment"], row["Action"], row.get("Genes", None)
if action.upper() == "ADD":
if padmetRef is None:
if verbose:
print("No given padmetRef, unable to copy %s" %element_id)
else:
if verbose:
print("Adding: %s" %(element_id))
padmetSpec.copyNode(padmetRef, element_id)
#reconstructionData:
if tool:
reconstructionData_id = element_id+"_reconstructionData_"+tool
reconstructionData = {"SOURCE": [source], "CATEGORY":[category], "TOOL":[tool], "COMMENT":[comment]}
if reconstructionData_id in list(padmetSpec.dicOfNode.keys()) and verbose:
print("Warning: The reaction %s seems to be already added from the same source %s" %(element_id, tool))
else:
reconstructionData_id = element_id+"_reconstructionData_MANUAL"
reconstructionData = {"SOURCE": [source], "CATEGORY":["MANUAL"], "COMMENT":[comment]}
if reconstructionData_id in list(padmetSpec.dicOfNode.keys()) and verbose:
print("Warning: The reaction %s seems to be already added from the same source 'MANUAL'" %element_id)
reconstructionData_rlt = Relation(element_id, "has_reconstructionData", reconstructionData_id)
padmetSpec.createNode("reconstructionData", reconstructionData_id, reconstructionData, [reconstructionData_rlt])
if genes_assoc:
#suppData:
if tool:
suppData_id = element_id+"_SuppData_"+tool
if suppData_id in list(padmetSpec.dicOfNode.keys()) and verbose:
print("Warning: The reaction %s seems to be already added from the same source %s" %(element_id, tool))
else:
suppData_id = element_id+"_SuppData_MANUAL"
if suppData_id in list(padmetSpec.dicOfNode.keys()) and verbose:
print("Warning: The reaction %s seems to be already added from the same source 'MANUAL'" %element_id)
suppData = {"GENE_ASSOCIATION":[genes_assoc]}
#create the node suppData and the relation has_suppData
suppData_rlt = Relation(element_id, "has_suppData", suppData_id)
padmetSpec.createNode("suppData", suppData_id, suppData, [suppData_rlt])
all_genes = parseGeneAssoc(genes_assoc)
nbGenes = len(all_genes)
if verbose:
print("%s is linked to %s genes" %(element_id, nbGenes))
for gene_id in all_genes:
try:
#check if gene already in the padmet
padmetSpec.dicOfNode[gene_id]
except KeyError:
padmetSpec.createNode("gene", gene_id)
#check if rxn already linked to gene x
try:
linked_rlt = [rlt for rlt in padmetSpec.dicOfRelationIn[element_id]
if rlt.type == "is_linked_to" and rlt.id_out == gene_id][0]
#rxn already linked to gene x, update misc
try:
linked_rlt.misc["SOURCE:ASSIGNMENT"].append(source)
except KeyError:
linked_rlt.misc["SOURCE:ASSIGNMENT"] = [source]
#rxn not linked to gene x
except IndexError:
linked_rlt = Relation(element_id, "is_linked_to", gene_id, {"SOURCE:ASSIGNMENT":[source]})
padmetSpec._addRelation(linked_rlt)
elif action.upper() == "DELETE":
if verbose:
print("deleting: %s" %(element_id))
padmetSpec.delNode(element_id)
elif action == "":
print("Nothing to do for: %s" %(element_id))
else:
print("Action: %s unknown for %s" %(action, element_id))
print("action must be = 'add' or 'delete' or ''")
exit()
padmetSpec.generateFile(output)
def rxn_creator(data_file, padmetSpec, output, padmetRef=None, source=None, tool=None, category="MANUAL", verbose=False):
"""
Read a data_file (form created with template_new_rxn and filed), for each reaction
to create, add the reaction in padmetSpec (only if the id of the reaction is not already in padmetSpec or in padmetRef if given)
the source ensure the traceability of the reaction, its a simple tag ex 'pathway_XX_update'
if not given the filename of data_file will be used.
if a tool was used to infer the reaction, define tool='name_of_the_tool'
the Padmet of reference padmetRef can be used to check that the reaction id is not
already in the database and copy information from the database for existing compounds
strongly recommended to give a padmetRef.
Parameters
----------
data_file: str
path to file based on template_new_rxn()
padmetSpec: padmet.classes.PadmetSpec
padmet to update
output: str
path to the new padmet file
source: str
tag associated to the new reactions to create and add, used for traceability
tool: str
The eventual tool used to infer the reactions to create and add
category: str
The default category of the reaction added manually is 'MANUAL'. Must not be changed.
padmetRef: padmet.classes.PadmetRef
padmet containing the database of reference
verbose: bool
if True print information
"""
if not source:
filename = os.path.splitext(os.path.basename(data_file))[0]
source = filename
source = source.upper()
if tool:
tool = tool.upper()
if not category:
category = "MANUAL"
dict_data = {}
with open(data_file, 'r') as f:
all_read = f.read()
sep = csv.Sniffer().sniff(all_read).delimiter
data = (line for line in all_read.splitlines() if len(line) != 0 and not line.startswith("#"))
for line in data:
#if len of value is 0 then TypeError raised
try:
attrib, value = line.split(sep)
except TypeError:
continue
attrib = attrib.replace(" ", "")
if attrib == "reaction_id":
current_id = value
dict_data[current_id] = {}
else:
try:
dict_data[current_id][attrib] .append(value)
except KeyError:
dict_data[current_id][attrib] = [value]
if verbose:
print("%s reactions to add" %len(list(dict_data.keys())))
for reaction_id, reaction_data in dict_data.items():
if verbose:
print("check if the id %s is already used" %reaction_id)
if reaction_id in list(padmetSpec.dicOfNode.keys()):
print("the id : %s is already associated to an other reaction in padmetSpec, choose an other" %reaction_id)
continue
if padmetRef is not None and reaction_id in list(padmetRef.dicOfNode.keys()):
print("the id : %s is already associated to an other reaction in padmetRef, choose an other" %reaction_id)
continue
if verbose:
print("Adding reaction %s" %reaction_id)
reaction_rev = reaction_data["reversible"][0].lower()
if reaction_rev.upper() == "TRUE":
reaction_rev = "REVERSIBLE"
elif reaction_rev.upper() == "FALSE":
reaction_rev = "LEFT-TO-RIGHT"
else:
print("Please choose a value in ['true','false'] for the reversibility of the reaction: %s" %reaction_id)
continue
comment = reaction_data["comment"]
node_misc = {"DIRECTION":[reaction_rev]}
padmetSpec.createNode("reaction", reaction_id, node_misc)
#reconstructionData:
if tool:
reconstructionData_id = reaction_id+"_reconstructionData_"+tool
reconstructionData = {"SOURCE": [source], "CATEGORY":[category], "TOOL":[tool], "COMMENT":comment}
if reconstructionData_id in list(padmetSpec.dicOfNode.keys()) and verbose:
print("Warning: The reaction %s seems to be already added from the same source %s" %(reaction_id, tool))
else:
reconstructionData_id = reaction_id+"_reconstructionData_MANUAL"
reconstructionData = {"SOURCE": [source], "CATEGORY":["MANUAL"], "COMMENT":comment}
if reconstructionData_id in list(padmetSpec.dicOfNode.keys()) and verbose:
print("Warning: The reaction %s seems to be already added from the same source 'MANUAL'" %reaction_id)
reconstructionData_rlt = Relation(reaction_id, "has_reconstructionData", reconstructionData_id)
padmetSpec.createNode("reconstructionData", reconstructionData_id, reconstructionData, [reconstructionData_rlt])
genes_assoc = reaction_data["linked_gene"][0]
if genes_assoc:
#suppData:
if tool:
suppData_id = reaction_id+"_SuppData_"+tool
if suppData_id in list(padmetSpec.dicOfNode.keys()) and verbose:
print("Warning: The reaction %s seems to be already added from the same source %s" %(reaction_id, tool))
else:
suppData_id = reaction_id+"_SuppData_MANUAL"
if suppData_id in list(padmetSpec.dicOfNode.keys()) and verbose:
print("Warning: The reaction %s seems to be already added from the same source 'MANUAL'" %reaction_id)
suppData = {"GENE_ASSOCIATION":[genes_assoc]}
#create the node suppData and the relation has_suppData
suppData_rlt = Relation(reaction_id, "has_suppData", suppData_id)
padmetSpec.createNode("suppData", suppData_id, suppData, [suppData_rlt])
all_genes = parseGeneAssoc(genes_assoc)
nbGenes = len(all_genes)
if verbose:
print("%s is linked to %s genes" %(reaction_id, nbGenes))
for gene_id in all_genes:
try:
#check if gene already in the padmet
padmetSpec.dicOfNode[gene_id]
except KeyError:
padmetSpec.createNode("gene", gene_id)
#check if rxn already linked to gene x
try:
linked_rlt = [rlt for rlt in padmetSpec.dicOfRelationIn[reaction_id]
if rlt.type == "is_linked_to" and rlt.id_out == gene_id][0]
#rxn already linked to gene x, update misc
try:
linked_rlt.misc["SOURCE:ASSIGNMENT"].append(source)
except KeyError:
linked_rlt.misc["SOURCE:ASSIGNMENT"] = [source]
#rxn not linked to gene x
except IndexError:
linked_rlt = Relation(reaction_id, "is_linked_to", gene_id, {"SOURCE:ASSIGNMENT":[source]})
padmetSpec._addRelation(linked_rlt)
if verbose:
print("check if all metabolites are already in the network")
try:
for reactant_data in reaction_data["reactant"]:
stoechio, metabo_id, compart = reactant_data.split(":")
stoechio = stoechio.replace(",", ".") #in case comma for sep
try:
padmetSpec.dicOfNode[metabo_id]
except KeyError:
if verbose:
print("%s not in the network" %metabo_id)
try:
if padmetRef is not None:
if verbose:
print("Try to copy from dbref")
padmetSpec._copyNodeExtend(padmetRef, metabo_id)
else:
raise KeyError
except KeyError:
if padmetRef is not None and verbose:
print("%s not in the padmetRef" %metabo_id)
if verbose:
print("creating a new compound")
padmetSpec.createNode("compound", metabo_id)
if verbose:
print(("new compound created: id = %s" %metabo_id))
rlt = Relation(reaction_id, "consumes", metabo_id)
rlt.misc.update({"STOICHIOMETRY":[stoechio], "COMPARTMENT":[compart]})
padmetSpec._addRelation(rlt)
except KeyError:
if verbose: print("No reactants defined")
try:
for product_data in reaction_data["product"]:
stoechio, metabo_id, compart = product_data.split(":")
stoechio = stoechio.replace(",", ".") #in case comma for sep
try:
padmetSpec.dicOfNode[metabo_id]
except KeyError:
if verbose:
print("%s not in the network" %metabo_id)
try:
if padmetRef is not None:
if verbose:
print("Try to copy from dbref")
padmetSpec._copyNodeExtend(padmetRef, metabo_id)
else:
raise KeyError
except KeyError:
if padmetRef is not None and verbose:
print("%s not in the padmetRef" %metabo_id)
if verbose:
print("creating a new compound")
padmetSpec.createNode("compound", metabo_id)
print("new compound created: id = %s" % metabo_id)
rlt = Relation(reaction_id, "produces", metabo_id)
rlt.misc.update({"STOICHIOMETRY":[stoechio], "COMPARTMENT":[compart]})
padmetSpec._addRelation(rlt)
except KeyError:
if verbose:
print("No products defined")
if "pathway" in reaction_data.keys():
pathways = reaction_data["pathway"][0].split(";")
for pwy_id in pathways:
try:
padmetSpec.dicOfNode[pwy_id]
except KeyError:
if verbose:
print("%s not in the network" %pwy_id)
if padmetRef is not None:
if verbose:
print("Check if new pathway %s is in dbref" %pwy_id)
if pwy_id in padmetRef.dicOfNode.keys():
print("Warning the new pathway %s exist in the dbref, risk of overwritting data, change pwy id" %pwy_id)
continue
padmetSpec.createNode("pathway", pwy_id)
if verbose:
print(("new pathway created: id = %s" %pwy_id))
rlt = Relation(reaction_id, "is_in_pathway", pwy_id)
padmetSpec._addRelation(rlt)
if verbose:
print("Creating output: %s" % output)
padmetSpec.generateFile(output)
def create_sbml_stat(species_name, sbml_file):
"""Extract reactions/pathways/compounds/genes from a sbml file.
Args:
species_name (str): species names
sbml_file (str): path to a sbml file
Returns
list: [species name, list of genes, list of reactions, list of reactions associated with genes, list of compounds]
"""
tree = etree.parse(sbml_file)
sbml = tree.getroot()
genes = []
reactions = []
gene_associated_rxns = []
fbc_gene_associated_rxns = []
fbc_rxn_associated_genes = []
compounds = []
for e in sbml:
if e.tag[0] == "{":
uri, tag = e.tag[1:].split("}")
else:
tag = e.tag
if tag == "model":
model_element = e
for els in model_element:
if 'listOfSpecies' in els.tag:
for el in els:
compounds.append(sbmlPlugin.convert_from_coded_id(el.get('metaid'))[0])
if 'listOfReactions' in els.tag:
for el in els:
reaction_id = sbmlPlugin.convert_from_coded_id(el.get('id'))[0]
reactions.append(reaction_id)
for subel in el.getchildren():
if 'notes' in subel.tag:
for subsubel in subel.getchildren():
for subsubsubel in subsubel.getchildren():
if 'GENE_ASSOCIATION' in subsubsubel.text:
for gene in sbmlPlugin.parseGeneAssoc(subsubsubel.text):
if gene not in genes:
genes.append(gene.replace('GENE_ASSOCIATION:', ''))
if reaction_id not in gene_associated_rxns:
gene_associated_rxns.append(reaction_id)
# Use geneProductAssociation for xml from MetaFlux.
elif 'geneProductAssociation' in subel.tag:
for subsubel in subel.getchildren():
if 'geneProductRef' in subsubel.tag:
gene = subsubel.get('{http://www.sbml.org/sbml/level3/version1/fbc/version2}geneProduct')
if gene:
gene = gene.replace('G_', '')
if gene not in fbc_rxn_associated_genes:
fbc_rxn_associated_genes.append(gene)
if reaction_id not in fbc_gene_associated_rxns:
fbc_gene_associated_rxns.append(reaction_id)
else:
for subsubsubel in subsubel.getchildren():
gene = subsubsubel.get('{http://www.sbml.org/sbml/level3/version1/fbc/version2}geneProduct')
if gene:
gene = gene.replace('G_', '')
if gene not in fbc_rxn_associated_genes:
fbc_rxn_associated_genes.append(gene)
if reaction_id not in fbc_gene_associated_rxns:
fbc_gene_associated_rxns.append(reaction_id)
# For XML from MetaFlux, use genes from geneProductAssociation to get genes and reaction with genes.
if len(genes) == 0:
if len(fbc_rxn_associated_genes) > 0:
genes = fbc_rxn_associated_genes
if len(gene_associated_rxns) == 0:
if len(fbc_gene_associated_rxns) > 0:
gene_associated_rxns = fbc_gene_associated_rxns
return [species_name, genes, reactions, gene_associated_rxns, compounds]
def enhance_db(metabolic_reactions, padmet, with_genes, verbose = False):
"""
Parse sbml metabolic_reactions and add reactions in padmet
if with_genes: add also genes information
Parameters
----------
metabolic_reactions: str
path to sbml metabolic-reactions.xml
padmet: padmet.PadmetRef
padmet instance
with_genes: bool
if true alos add genes information.
Returns
-------
padmet.padmetRef:
padmet instance with pgdb within pgdb + metabolic-reactions.xml data
"""
print("loading sbml file: %s" %metabolic_reactions)
reader = libsbml.SBMLReader()
document = reader.readSBML(metabolic_reactions)
for i in range(document.getNumErrors()):
print(document.getError(i).getMessage())
model = document.getModel()
listOfReactions = model.getListOfReactions()
#recovere the reactions that are not in the basic metacyc but in the sbml file
#use the reactions_name instead of ids because the ids are encoded, the name is the non-encoded version of the id
padmet_reactions_id = set([node.id for node in list(padmet.dicOfNode.values()) if node.type == "reaction"])
reaction_to_add = [reaction for reaction in listOfReactions
if reaction.getName() not in padmet_reactions_id]
count = 0
if verbose: print(str(len(reaction_to_add))+" reactions to add")
for reactionSBML in reaction_to_add:
count += 1
reaction_id = reactionSBML.getName()
if verbose: print(str(count)+"/"+str(len(reaction_to_add))+"\t"+reaction_id)
if reactionSBML.getReversible():
reaction_dir = "REVERSIBLE"
else:
reaction_dir = "LEFT-TO-RIGHT"
try:
reaction_node = padmet.dicOfNode[reaction_id]
except KeyError:
reaction_node = Node("reaction", reaction_id, {"DIRECTION": [reaction_dir]})
padmet.dicOfNode[reaction_id] = reaction_node
reactants = reactionSBML.getListOfReactants()
for reactant in reactants: #convert ids
reactant_id, _type, reactant_compart = sbmlPlugin.convert_from_coded_id(reactant.getSpecies())
if reactant_id not in list(padmet.dicOfNode.keys()):
reactant_node = Node("compound",reactant_id)
padmet.dicOfNode[reaction_id] = reactant_node
reactant_stoich = reactant.getStoichiometry()
consumes_rlt = Relation(reaction_id,"consumes",reactant_id, {"STOICHIOMETRY":[reactant_stoich], "COMPARTMENT": [reactant_compart]})
list_of_relation.append(consumes_rlt)
products = reactionSBML.getListOfProducts()
for product in products:
product_id, _type, product_compart = sbmlPlugin.convert_from_coded_id(product.getSpecies())
if product_id not in list(padmet.dicOfNode.keys()):
product_node = Node("compound",product_id)
padmet.dicOfNode[product_id] = product_node
product_stoich = product.getStoichiometry()
produces_rlt = Relation(reaction_id,"produces",product_id,{"STOICHIOMETRY": [product_stoich], "COMPARTMENT": [product_compart]})
list_of_relation.append(produces_rlt)
if with_genes:
notes = sbmlPlugin.parseNotes(reactionSBML)
if "GENE_ASSOCIATION" in list(notes.keys()):
#Using sbmlPlugin to recover all genes associated to the reaction
listOfGenes = sbmlPlugin.parseGeneAssoc(notes["GENE_ASSOCIATION"][0])
if len(listOfGenes) != 0:
for gene in listOfGenes:
try:
#check if gene already in the padmet
padmet.dicOfNode[gene]
except TypeError:
gene_node = Node("gene",gene)
padmet.dicOfNode[gene] = gene_node
is_linked_rlt = Relation(reaction_id, "is_linked_to", gene)
list_of_relation.append(is_linked_rlt)
return padmet
