network = read_csv('%s/files/string_mouse_network_filtered_800.txt' % wd, sep='\t')
network_proteins = set(network['protein1']).intersection(network['protein2'])
# ---- Set-up UniProt
uniprot = UniProt(cache=True)
# ---- Set-up QuickGO bioservice
quickgo = QuickGO(cache=True)
# ---- Set-up KEGG bioservice
kegg, kegg_parser = KEGG(cache=True), KEGGParser()
kegg.organism = 'mmu'
print '[INFO] KEGG service configured'
kegg_pathways = {p: kegg.parse_kgml_pathway(p) for p in kegg.pathwayIds}
print '[INFO] KEGG pathways extracted: ', len(kegg_pathways)
# Convert KEGG pathways Gene Name to UniProt
k2u = kegg.conv('uniprot', 'mmu')
kegg_pathways_proteins = {p: {k2u[x].split(':')[1] for i in kegg_pathways[p]['entries'] if i['type'] == 'gene' for x in i['name'].split(' ') if x in k2u} for p in kegg_pathways}
kegg_uniprot_acc_map = {x for p in kegg_pathways_proteins for x in kegg_pathways_proteins[p]}
kegg_uniprot_acc_map = {p: uniprot.get_fasta(str(p)).split(' ')[0].split('|')[2] for p in kegg_uniprot_acc_map}
kegg_pathways_proteins = {p: {kegg_uniprot_acc_map[i] for i in kegg_pathways_proteins[p]} for p in kegg_pathways_proteins}
print '[INFO] KEGG pathways Ids converted to UniProt: ', len(kegg_pathways_proteins)
# ---- Set-up GO Terms gene list
go_terms_file = '%s/files/go_terms_uniprot.pickle' % wd
def pathway_to_dataframe(pathway_id, org='hsa', verbose=False, cache=False):
"""
Extract protein-protein interaction from KEGG pathway to
a pandas DataFrame. NOTE: Interactions will be directionless.
Parameters
----------
pathway_id : str
Pathway identifier to parse into a dataframe. Example: 'path:hsa00010'
org : str or None, optioanl, default: 'hsa'
If supplied, filters out all interactions with identifiers that
are not in the dictionary created from :func:`kegg_to_uniprot`. If None,
all interactions are parsed regardless of mappability to UniProt.
verbose : bool, optional, default: False
If True, logs messages to stdout to inform of current progress.
cache : bool, optional, default: False
If True, HTTP responses are cached by `bioservices`. This can save
time but you will eventually miss out on new database releases if
your cache is old.
Returns
-------
`pd.DataFrame`
DataFrame with 'source', 'target', 'label', 'pubmed', and
'experiment_type' columns.
"""
kegg = KEGG(cache=cache)
kegg.organism = org
kegg_to_up = kegg_to_uniprot(org, cache)
res = kegg.parse_kgml_pathway(pathway_id)
sources = []
targets = []
labels = []
if verbose:
logger.info("Parsing pathway {}".format(pathway_id))
for rel in res['relations']:
id1 = rel['entry1']
id2 = rel['entry2']
name1 = res['entries'][[x['id']
for x in res['entries']].index(id1)]['name']
name2 = res['entries'][[x['id']
for x in res['entries']].index(id2)]['name']
type1 = res['entries'][[x['id']
for x in res['entries']].index(id1)]['type']
type2 = res['entries'][[x['id']
for x in res['entries']].index(id2)]['type']
reaction_type = rel['name'].replace(' ', '-')
link_type = rel['link']
if link_type not in links_to_include:
continue
if type1 not in types_to_include or type2 not in types_to_include:
continue
for a in name1.strip().split(' '):
for b in name2.strip().split(' '):
valid_db_a = (kegg.organism in a or 'ec' in a)
valid_db_b = (kegg.organism in b or 'ec' in b)
valid_db_a &= (a in kegg_to_up)
valid_db_b &= (b in kegg_to_up)
if valid_db_a and valid_db_b:
sources.append(a)
targets.append(b)
labels.append(reaction_type)
interactions = make_interaction_frame(sources, targets, labels)
return interactions
result = {}
for inputGene in inputGenes:
print('Processing gene ' + inputGene + ':')
try:
pathways = k.get_pathway_by_gene(inputGene, 'hsa')
except (AttributeError):
print('Invalid gene identifier')
continue
else:
if pathways:
for pathway in pathways:
if pathway != 'hsa01100':
print('\tProcessing pathway ' + pathway)
rel = k.parse_kgml_pathway(pathway)
genes_result = []
####### Part where the functions are called
for entry in rel['entries']:
if entry['type'] == 'gene':
for aGene in getGeneNameFromEntries(entry['name']):
if aGene == inputGene:
getTargetsFromRelations(rel, entry)
############################
unique_genes = list(set(genes_result))
if pathway in result:
result[pathway] += genes_result
else: