def test_read_and_write_KGML_files(self):
"""Read KGML from, and write KGML to, local files.
Check we read/write the correct number of elements.
"""
for p in self.data:
# Test opening file
with open(p.infilename) as f:
pathway = read(f)
# Do we have the correct number of elements of each type
self.assertEqual(
(
len(pathway.entries),
len(pathway.orthologs),
len(pathway.compounds),
len(pathway.maps),
),
p.element_counts,
)
# Test writing file
with open(p.outfilename, "w") as f:
f.write(pathway.get_KGML())
# Can we read the file we wrote?
with open(p.outfilename) as f:
pathway = read(f)
# Do we have the correct number of elements of each type
self.assertEqual(
(
len(pathway.entries),
len(pathway.orthologs),
len(pathway.compounds),
len(pathway.maps),
),
p.element_counts,
)
def test_read_and_write_KGML_files(self):
"""Read KGML from, and write KGML to, local files.
Check we read/write the correct number of elements.
"""
for p in self.data:
# Test opening file
with open(p.infilename, 'rU') as f:
pathway = read(f)
# Do we have the correct number of elements of each type
self.assertEqual((len(pathway.entries),
len(pathway.orthologs),
len(pathway.compounds),
len(pathway.maps)),
p.element_counts)
# Test writing file
with open(p.outfilename, 'w') as f:
f.write(pathway.get_KGML())
# Can we read the file we wrote?
with open(p.outfilename, 'rU') as f:
pathway = read(f)
# Do we have the correct number of elements of each type
self.assertEqual((len(pathway.entries),
len(pathway.orthologs),
len(pathway.compounds),
len(pathway.maps)),
p.element_counts)
def test_render_KGML_modify(self):
"""Rendering of KGML to PDF, with modification."""
# We test rendering of the original KGML for KO01100,
# modifying line width for the lipid pathway
p = self.data
with open(p[0].infilename) as f:
pathway = read(f)
mod_rs = [
e for e in pathway.orthologs
if len(set(e.name.split()).intersection(self.ko_ids))
]
for r in mod_rs:
for g in r.graphics:
g.width = 10
kgml_map = KGMLCanvas(pathway)
kgml_map.draw(p[0].output_stem + '_widths.pdf')
# We test rendering of the original KGML for KO3070,
# modifying the reaction colours for each ortholog entry
with open(p[1].infilename) as f:
pathway = read(f)
orthologs = [e for e in pathway.orthologs]
# Use Biopython's ColorSpiral to generate colours
cs = ColorSpiral(a=2, b=0.2, v_init=0.85, v_final=0.5, jitter=0.03)
colors = cs.get_colors(len(orthologs))
for o, c in zip(orthologs, colors):
for g in o.graphics:
g.bgcolor = c
kgml_map = KGMLCanvas(pathway)
pathway.image = p[1].pathway_image
kgml_map.import_imagemap = p[1].show_pathway_image
kgml_map.draw(p[1].output_stem + '_colors.pdf')
def test_render_KGML_modify(self):
""" Rendering of KGML to PDF, with modification.
"""
# We test rendering of the original KGML for KO01100,
# modifying line width for the lipid pathway
p = self.data
with open(p[0].infilename) as f:
pathway = read(f)
mod_rs = [e for e in pathway.orthologs if
len(set(e.name.split()).intersection(self.ko_ids))]
for r in mod_rs:
for g in r.graphics:
g.width = 10
kgml_map = KGMLCanvas(pathway)
kgml_map.draw(p[0].output_stem + '_widths.pdf')
# We test rendering of the original KGML for KO3070,
# modifying the reaction colours for each ortholog entry
with open(p[1].infilename) as f:
pathway = read(f)
orthologs = [e for e in pathway.orthologs]
# Use Biopython's ColorSpiral to generate colours
cs = ColorSpiral(a=2, b=0.2, v_init=0.85, v_final=0.5,
jitter=0.03)
colors = cs.get_colors(len(orthologs))
for o, c in zip(orthologs, colors):
for g in o.graphics:
g.bgcolor = c
kgml_map = KGMLCanvas(pathway)
pathway.image = p[1].pathway_image
kgml_map.import_imagemap = p[1].show_pathway_image
kgml_map.draw(p[1].output_stem + '_colors.pdf')
def test_render_KGML_transparency(self):
"""Rendering of KGML to PDF, with color alpha channel."""
# We test rendering of the original KGML for KO01100,
# modifying alpha channel for the lipid pathway
p = self.data
with open(p[0].infilename) as f:
pathway = read(f)
mod_rs = [e for e in pathway.orthologs if
len(set(e.name.split()).intersection(self.ko_ids))]
for r in mod_rs:
for g in r.graphics:
# Modify hex colour directly by appending alpha channel
# to hex string
g.fgcolor = g.fgcolor + "77"
g.width = 20
kgml_map = KGMLCanvas(pathway)
kgml_map.draw(p[0].output_stem + "_transparency.pdf")
# We test rendering of the original KGML for KO3070,
# modifying the alpha channel for each ortholog entry
with open(p[1].infilename) as f:
pathway = read(f)
orthologs = list(pathway.orthologs)
# Use Biopython's ColorSpiral to generate colours
cs = ColorSpiral(a=2, b=0.2, v_init=0.85, v_final=0.5,
jitter=0.03)
colors = cs.get_colors(len(orthologs))
for o, c in zip(orthologs, colors):
# Modify color tuples to add alpha channel
c = c + (0.5, )
for g in o.graphics:
g.bgcolor = c
kgml_map = KGMLCanvas(pathway)
pathway.image = p[1].pathway_image
kgml_map.import_imagemap = p[1].show_pathway_image
kgml_map.draw(p[1].output_stem + "_transparency.pdf")
def test_render_KGML_transparency(self):
"""Rendering of KGML to PDF, with color alpha channel."""
# We test rendering of the original KGML for KO01100,
# modifying alpha channel for the lipid pathway
p = self.data
with open(p[0].infilename) as f:
pathway = read(f)
mod_rs = [e for e in pathway.orthologs if
len(set(e.name.split()).intersection(self.ko_ids))]
for r in mod_rs:
for g in r.graphics:
# Modify hex colour directly by appending alpha channel
# to hex string
g.fgcolor = g.fgcolor + "77"
g.width = 20
kgml_map = KGMLCanvas(pathway)
kgml_map.draw(p[0].output_stem + '_transparency.pdf')
# We test rendering of the original KGML for KO3070,
# modifying the alpha channel for each ortholog entry
with open(p[1].infilename) as f:
pathway = read(f)
orthologs = [e for e in pathway.orthologs]
# Use Biopython's ColorSpiral to generate colours
cs = ColorSpiral(a=2, b=0.2, v_init=0.85, v_final=0.5,
jitter=0.03)
colors = cs.get_colors(len(orthologs))
for o, c in zip(orthologs, colors):
# Modify color tuples to add alpha channel
c = c + (0.5, )
for g in o.graphics:
g.bgcolor = c
kgml_map = KGMLCanvas(pathway)
pathway.image = p[1].pathway_image
kgml_map.import_imagemap = p[1].show_pathway_image
kgml_map.draw(p[1].output_stem + '_transparency.pdf')
def getPathway(pathName):
try:
response = requests.get('http://rest.kegg.jp/get/' + pathName +
'/kgml')
except requests.exceptions.HTTPError as err:
raise SystemExit(err)
pathway = read(response.text)
return pathway
def test_render_KGML_basic(self):
"""Basic rendering of KGML: write to PDF without modification."""
# We test rendering of the original KEGG KGML using only local
# files.
for p in self.data:
with open(p.infilename, 'rU') as f:
pathway = read(f)
pathway.image = p.pathway_image
kgml_map = KGMLCanvas(pathway)
kgml_map.import_imagemap = p.show_pathway_image
kgml_map.draw(p.output_stem + '_original.pdf')
def test_render_KGML_basic(self):
"""Basic rendering of KGML: write to PDF without modification."""
# We test rendering of the original KEGG KGML using only local
# files.
for p in self.data:
with open(p.infilename, 'rU') as f:
pathway = read(f)
pathway.image = p.pathway_image
kgml_map = KGMLCanvas(pathway)
kgml_map.import_imagemap = p.show_pathway_image
kgml_map.draw(p.output_stem + '_original.pdf')
def test_render_KGML_import_map(self):
"""Basic rendering of KGML: use imported imagemap.
Uses the URL indicated in the .xml file.
This test may fail if the imagemap is not available (e.g. if
there is not a web connection), and may look odd if the remote
imagemap has changed since the local KGML file was downloaded.
"""
# We test rendering of the original KEGG KGML using imported files
for p in self.data:
with open(p.infilename) as f:
pathway = read(f)
kgml_map = KGMLCanvas(pathway, import_imagemap=True)
kgml_map.draw(p.output_stem + "_importmap.pdf")
def test_render_KGML_import_map(self):
"""Basic rendering of KGML: use imported imagemap
Uses the URL indicated in the .xml file.
This test may fail if the imagemap is not available (e.g. if
there is not a web connection), and may look odd if the remote
imagemap has changed since the local KGML file was downloaded.
"""
# We test rendering of the original KEGG KGML using imported files
for p in self.data:
with open(p.infilename, 'rU') as f:
pathway = read(f)
kgml_map = KGMLCanvas(pathway, import_imagemap=True)
kgml_map.draw(p.output_stem + '_importmap.pdf')
def parse(self, file):
def _set_gene(way, path):
gene_name = [
e.name for e in way.entries.values() if e.type == 'gene'
]
ent_id = [e.id for e in way.entries.values() if e.type == 'gene']
name = _parse_name(gene_name)
for i in range(len(name)):
path[ent_id[i]] = Node(name[i])
def _parse_name(nodes):
names = []
for node in nodes:
names.append(node.split())
for i in range(0, len(names)):
temp = []
for name in names[i]:
temp.append(name.split(':')[1])
names[i] = temp
return names
def _set_weight(way, path):
for relation in way.relations:
if relation.entry1.type == 'gene':
path[relation.entry1.id].weight += 1
if relation.entry2.type == 'gene':
path[relation.entry2.id].weight += 1
def _change_keys(path):
i = 0
for key in list(path.keys()):
path[i] = path.pop(key)
i = i + 1
f = open('D:/NIR/bernmix_method/hsaFiles/' + file)
pathway = read(f, 'r')
_set_gene(pathway, self.path)
_set_weight(pathway, self.path)
_change_keys(self.path)
def init(self):
with open(self.ko_list_path) as h:
for koline in h.readlines():
ko, desc = koline.replace("#160;", "").split("\t")
gene = desc.split(";")[0]
desc2 = ";".join(desc.split(";")[1:])
desc = desc2.strip().split("[EC:")[0]
ecs = None
if len(desc2.split("[EC:")) > 1:
ecs = desc2.split("[EC:")[1].replace("]", "")
ecs = ["ec:" + x for x in ecs.split()]
self.ko_dict[ko]["genes"] = gene.split()
self.ko_dict[ko]["ecs"] = ecs
self.ko_dict[ko]["desc"] = desc
self.ko_dict[ko]["pathways"] = []
for kgml_path in glob.glob(self.kgmls_dir + "/*.kgml"):
if os.path.getsize(kgml_path) > 100:
kgml = read(open(kgml_path))
pw_name = "kegg_pw:" + kgml.name.split(":ko")[1]
ko_react = defaultdict(lambda: [])
for ortholog in kgml.orthologs:
for ko in ortholog.name.split():
if ":" in ko and "pathways" in self.ko_dict[ko]:
self.ko_dict[ko]["pathways"].append(pw_name)
if ortholog.reaction:
ko_react[ko.split(":")[1]].append(
ortholog.reaction)
self.pw_dict[pw_name] = {
"name": pw_name,
"title": kgml.title,
"reactions": dict(ko_react)
}
self.ko_dict = dict(self.ko_dict)
def __init__(self, filePath):
self.compoundsGraph = {}
self.bioKGML = read(open(filePath, 'r'))
self._fillGraph()
def parse_kegg(org_id, strain, sourcedb, session):
# get pathways for organism specified by org_id
pathways = kegg_list(database='pathway', org=org_id).read().split('path:')
path_ids = []
# make list of path ids to iterate through
for path in pathways:
if path != '':
path_ids.append(path[:8])
# iterate through each path and obtain interactions
for path in path_ids:
# get kgml representation of path
kgml_path = read(kegg_get(path, option='kgml'))
path_name = kgml_path._getname()
# dictionary of compounds in current path (node_id: kegg_id)
# compound._getid() returns node id (only relevant in context of current path)
# compound._getname() returns kegg id (relevant in overall KEGG DB)
compound_ids = {}
for compound in kgml_path.compounds:
compound_ids[compound._getid()] = compound._getname()[-6:]
# go through each relation in path
for relation in kgml_path.relations:
relation_type = relation.element.attrib['type']
# ignore maplink relations
if relation_type == 'maplink': continue
# relation._getentry1/2() returns protein id (locus) or compound id (KEGG id)
entries = [relation._getentry1()._getname(), relation._getentry2()._getname()]
# if one or both interactors are listed as undefined, move on to next interaction
if (entries[0] == 'undefined') | (entries[1] == 'undefined'): continue
# list to hold existing interactors
interactors = [[], []]
# list to hold new metabolite ids for interactions with metabolites not yet in the database
new_metabolites = [[], []]
# go through each entry in the relation
for num in range(0, 2):
# each entry may contain >1 id; go through all of them
for id in entries[num].split(' '):
if id == '': continue
# if interactor is not protein or compound, continue
if (id.split(':')[0] != org_id) & (id.split(':')[1] not in kegg_compounds): continue
# check if the id is a kegg id by searching in kegg_compounds
kegg_id= None
if id.split(':')[1] in kegg_compounds:
kegg_id = id.split(':')[1]
# check if interactor (protein) already exists
if (kegg_id is None) & (org_id != 'eco'):
interactor = session.query(Interactor).get(id.split(':')[1])
if interactor is not None:
# make sure to add None value; this will be needed to create interaction reference later
# None is appended rather than the interactor id because the interactor is not an ortholog
interactors[num].append([interactor, None])
# if it doesnt exist, it's not a valid protein, so check if it is a valid compound
elif kegg_id is not None:
interactor = session.query(Metabolite).filter_by(kegg = kegg_id).first()
# if metabolite with id was not found, append the kegg_id to new_metabolites to create
if interactor is None:
new_metabolites[num].append(kegg_id)
else:
# if the metabolite was found, add it to the existing interactor list
interactors[num].append([interactor, interactor.id])
# if parsing E. coli path, add all orthologs to interactor list
elif org_id == 'eco':
for ortholog in session.query(OrthologEcoli).filter_by(ortholog_id = id.split(':')[1],
strain_protein = strain).all():
if ortholog is not None:
# add the id of the ecoli protein for the interaction reference later
interactors[num].append([ortholog.protein, id.split(':')[1]])
# create list of interactor pairs from two separate lists
interactor_pairs = []
# create interactor pairs from interactors which already exist in db
for interactor1 in interactors[0]:
for interactor2 in interactors[1]:
if (interactor1[0].type != 'm') | (interactor2[0].type != 'm'):
interactor_pairs.append([interactor1, interactor2])
# create interactor pair from interactors and new metabolites
for interactor1 in interactors[0]:
for id in new_metabolites[1]:
# ignore interactor pairs which would result in m-m interactions
if interactor1[0].type == 'm': continue
# Note: can query metabolite with kegg only because we updated the metabolite info first
metabolite = session.query(Metabolite).filter_by(kegg = id).first()
if metabolite is None:
metabolite = Metabolite(id = id, kegg = id, pubchem = kegg_compounds[id]['pubchem'],
chebi = kegg_compounds[id]['chebi'])
session.add(metabolite)
interactor_pairs.append([interactor1, [metabolite, metabolite.id]])
for interactor1 in interactors[1]:
for id in new_metabolites[0]:
if interactor1[0].type == 'm': continue
metabolite = session.query(Metabolite).filter_by(kegg = id).first()
if metabolite is None:
metabolite = Metabolite(id = id, kegg = id, pubchem = kegg_compounds[id]['pubchem'],
chebi = kegg_compounds[id]['chebi'])
session.add(metabolite)
interactor_pairs.append([interactor1, [metabolite, metabolite.id]])
# if no interactor pairs were found, move on the the next interaction
if len(interactor_pairs) == 0: continue
# get all intermediates in reaction of type compound
intermeds = []
for subtype in relation.element.iter(tag='subtype'):
# if the subtype element is a compound, get its node id
if 'compound' in subtype.attrib:
compound_node_id = subtype.attrib['compound']
if compound_node_id is None: continue
# if the node id was not stored in the compound ids for this path, move on to the next sybtype
if int(compound_node_id) not in compound_ids: continue
# if compound id is valid, either add existing matching metabolite or create new one and add
kegg_id = compound_ids[int(compound_node_id)]
metabolite = session.query(Metabolite).filter_by(kegg = kegg_id).first()
if metabolite is None:
metabolite = Metabolite(id=kegg_id, name=kegg_compounds[kegg_id]['name'],
pubchem=kegg_compounds[kegg_id]['pubchem'],
chebi=kegg_compounds[kegg_id]['chebi'], kegg=kegg_id)
session.add(metabolite)
intermeds.append([metabolite, metabolite.id])
# add protein - intermediate interactor pairs
for interactor_list in interactors:
for interactor in interactor_list:
if interactor[0].type != 'm':
for intermed in intermeds:
interactor_pairs.append([interactor, intermed])
# go through each interaction pair and add interaction if it doesnt exist yet
for interactor_pair in interactor_pairs:
homogenous = (interactor_pair[0][0] == interactor_pair[1][0])
interaction = session.query(Interaction).filter(Interaction.interactors.contains(interactor_pair[0][0]),
Interaction.interactors.contains(interactor_pair[1][0]),
Interaction.homogenous == homogenous).first()
source = session.query(InteractionSource).filter_by(data_source=sourcedb).first()
#create interaction if it doesnt exist yet, add source to its sources if it isn't already
if interaction is None:
interaction = Interaction(type=interactor_pair[0][0].type + '-' + interactor_pair[1][0].type,
strain=strain, homogenous=homogenous,
interactors=[interactor_pair[0][0], interactor_pair[1][0]])
interaction.sources.append(source)
if org_id == 'eco':
interaction.ortholog_derived = 'Ecoli'
session.add(interaction), session.commit()
elif source not in interaction.sources:
interaction.sources.append(source)
# in case the interaction already existed, make sure interactor_a and interactor_b variables for
# new interaction reference match up with the first and second interactors of the existing
# interaction (so it's easy to see which Pseudomonas interactor matches up with which Ecoli
# ortholog if the org id is eco)
interactor_a, interactor_b = None, None
if org_id == 'eco':
if interaction.interactors[0] == interactor_pair[0][0]:
interactor_a = interactor_pair[0][1]
interactor_b = interactor_pair[1][1]
else:
interactor_b = interactor_pair[0][1]
interactor_a = interactor_pair[1][1]
# search for reference
reference = session.query(InteractionReference).filter_by(source_db='kegg',
comment='in ' + path_name + ' path',
interactor_a=interactor_a,
interactor_b=interactor_b).first()
# if the reference doesnt exist, create it, add it to the interaction's references and add the source
# to the reference's sources
if reference is None:
reference = InteractionReference(source_db='kegg', comment='in ' + path_name + ' path',
interactor_a=interactor_a, interactor_b=interactor_b)
interaction.references.append(reference)
reference.sources.append(source)
# if the reference does exist, add it to the interaction's reference list and add the source to the
# reference's source list if it isn't there already
else:
if interaction not in reference.interactions:
reference.interactions.append(interaction)
if source not in reference.sources:
reference.sources.append(source)
session.commit()
print(sourcedb, session.query(Interaction).count())
def pathwayId2PathwayInfo(pid):
url2 = 'curl http://rest.kegg.jp/get/' + pid + '/kgml'
out = os.system(url2 + "> pathway.xml")
pathway = read(open("pathway.xml"))
return pathway
