def teste4():
s = KEGG()
s.organism = "hsa" #H**o sapiens (human)
modules=s.moduleIds #pathway modules
dic=s.parse(s.get(modules[0]))
compounds=dic["COMPOUND"]#dictionary with the names of the compounds {'C00074': 'Phosphoenolpyruvate',.....
pathway=dic["PATHWAY"] # {'map00010': 'Glycolysis / Gluconeogenesis',......
module_name=dic["NAME"] #['Glycolysis (Embden-Meyerhof pathway), glucose => pyruvate']
return pathway
def __init__(self, gene_lists, taxon, dataframe,
kegg_organism=None,
enrichment_params={
"padj": 0.05,
"log2_fc": 3,
"max_entries": 3000,
"kegg_background": None,
"mapper": None,
"preload_directory": None,
'plot_compute_levels': False,
'plot_logx': True
},
go_only=False,
kegg_only=False,
command=""
):
""".. rubric:: constructor
"""
super().__init__()
self.title = "Enrichment"
self.command = command
#self.rnadiff_folder = rnadiff_folder
self.gene_lists = gene_lists
self.enrichment_params = enrichment_params
self.data = dataframe
self.taxon = taxon
if taxon == 10090:
self.organism = "mmu"
elif taxon == 9606:
self.organism = "hsa"
else:
if kegg_organism is None:
logger.error("You must specify the kegg organism name if not human or mouse: eg., eco for ecoli")
# figure out the organism from taxon
raise NotImplementedError
else:
from bioservices import KEGG
k = KEGG()
k.organism = kegg_organism # validates the organism name
self.organism = kegg_organism
if self.enrichment_params['preload_directory']:
pathname = self.enrichment_params['preload_directory']
if os.path.exists(pathname) is False:
logger.error(f"{pathname} does not exist")
sys.exit(1)
#from sequana.rnadiff import RNADiffResults
#self.rnadiff = RNADiffResults(self.rnadiff_folder)
self.rnadiff = {}
self.create_report_content(go_only=go_only, kegg_only=kegg_only)
self.create_html("enrichment.html")
def teste5():
s = KEGG()
s.organism = "hsa" #H**o sapiens (human)
modules=s.moduleIds #pathway modules
dic=s.parse(s.get("M00627"))
module_name=dic["NAME"][0]
reactions=dic["REACTION"]
if "Pentose phosphate cycle" in module_name:
print(module_name)
else:
print("haha")
def teste2():
s = KEGG()
s.organism = "hsa"
modules=s.moduleIds
print(modules[3])
dic=s.parse(s.get(modules[3]))
reactions=dic["REACTION"]
dic_reac={}
for reac in reactions:
teste=reactions[reac]
string=teste.split(" ")
dic_reac[reac]=string
return dic_reac #it gives a dictionary with reactionsID as keys and a list of compounds
def teste6():
s = KEGG()
s.organism = "hsa"
modules=["M00001", "M00002", "M00013", "M00034"]
dic_reac={}
for mod in modules:
dic=s.parse(s.get(mod))
reactions=dic["REACTION"]
for reac in reactions:
teste=reactions[reac]
string=teste.split(" ")
dic_reac[reac]=string
return dic_reac
def download_pathway_ids(organism, cache=False):
"""
Query KEGG for a recent list of pathways for an organism.
Parameters
----------
organism: str
A KEGG organism code. For example 'hsa'.
cache : bool, optional, default: False
If True, results are cached by `bioservices`. This can save
time but you will eventually miss out on new database releases if
your cache is old.
Returns
-------
`list`
List of str pathway identifiers.
"""
kegg = KEGG(cache=cache)
kegg.organism = organism
pathways = kegg.pathwayIds
return pathways
def kegg():
k = KEGG()
k.organismIds
k.organism = "hsa"
return k
def find_pathways_organism(cvDict,
preDefList=[],
writeGraphml=True,
organism="hsa"):
aliasDict, koDict, orgDict = {}, {}, {
} # set up empty dictionaries for converting codes
nc.parseKEGGdict('inputData/ko00001.keg', aliasDict,
koDict) # parse the dictionary of ko codes
try: # try to retrieve and parse the dictionary containing organism gene names to codes conversion
url = urllib2.urlopen('http://rest.kegg.jp/list/' + organism)
text = url.readlines()
# reads KEGG dictionary of identifiers between numbers and actual protein names and saves it to a python dictionary
for line in text:
line_split = line.split('\t')
k = line_split[0].split(':')[1]
nameline = line_split[1].split(';')
name = nameline[0]
if ',' in name:
nameline = name.split(',')
name = nameline[0]
for entry in range(1, len(nameline)):
aliasDict[nameline[entry].strip()] = name.upper()
orgDict[k] = name
except:
print('Could not get library: ' + organism)
k = KEGG() # read KEGG from bioservices
k.organism = organism
minOverlap = 5
if len(preDefList) == 0:
pathwayList = list(k.pathwayIds)
else:
pathwayList = list(preDefList)
# set up a converter to retain only numbers from KEGG pathway codes
allChars = string.maketrans('', '')
noDigits = allChars.translate(allChars, string.digits)
genes = set(cvDict.keys()) # find the list of genes included in dataset
for x in pathwayList:
x = x.replace("path:", "")
code = str(x)
code = code.translate(allChars, noDigits) # eliminate org letters
coder = str('ko' + code) # add ko
graph = nx.DiGraph() # open a graph object
nc.uploadKEGGcodes([coder], graph, koDict) # get ko pathway
coder = str(organism + code) # set up with org letters
uploadKEGGcodes_org([coder], graph, orgDict, koDict,
organism) # get org pathway
# check to see if there is a connected component, simplify graph and print if so
allNodes = set(graph.nodes())
test = len(allNodes.intersection(genes))
print("Pathway: ", x, " Overlap: ", test, " Edges: ",
len(graph.edges()))
if len(
list(nx.connected_component_subgraphs(graph.to_undirected()))
) > 0: # if there is more than a 1 node connected component, run BONITA
#nx.write_graphml(graph,coder+'_before.graphml')
if len(genes.intersection(graph.nodes())
) > minOverlap: # if there are 5 genes shared
graph = simplifyNetworkpathwayAnalysis(
graph, cvDict) # simplify graph to nodes in dataset
nx.write_graphml(graph, coder + '.graphml') # write graph out
nx.write_gpickle(graph, coder + '.gpickle') # write graph out
print('nodes: ', str(len(graph.nodes())), ', edges:',
str(len(graph.edges())))
print(graph.nodes())
if len(graph.nodes()) > 0:
# save the removed nodes and omics data values for just those nodes in the particular pathway
pathwaySampleList = [
{}
for q in range(len(geneDict[list(graph.nodes())[0]]))
]
for noder in graph.nodes():
for jn in range(len(pathwaySampleList)):
pathwaySampleList[jn][noder] = geneDict[noder][jn]
pickle.dump(pathwaySampleList,
open(coder + "_sss.pickle", "wb"))
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
sns.set(style='ticks', palette='pastel', color_codes=True)
# ---- Import network
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)
def kegg():
k = KEGG()
k.organismIds
k.organism = "hsa"
return k
