def plot_all():
paths = (hil_neg, hil_pos, syn_neg)
metlin = []
hmdb = []
metfrag = []
for path in paths:
metlin.append(read.metlin(path))
hmdb.append(read.hmdb(path))
metfrag.append(
set(read.metfrag_with_scores(path, keep_zero_scores=False).keys()))
# remove non-features & cofactors
for i in range(len(paths)):
metlin[i] = (metlin[i] & features) - cofactors
hmdb[i] = (hmdb[i] & features) - cofactors
metfrag[i] = (metfrag[i] & features) - cofactors
all = []
for db in [metlin, hmdb, metfrag]:
all.append(reduce(lambda x, y: x | y, db))
figure, axes = plt.subplots(2, 3)
venn3(all, ('metlin', 'hmdb', 'metfrag'), ax=axes[0][0])
venn3((all[0] | all[1], all[2], features),
('observed', 'metfrag', 'features'),
ax=axes[0][1])
venn3(metlin, ('hilneg', 'hilpos', 'synneg'), ax=axes[1][0])
venn3(hmdb, ('hilneg', 'hilpos', 'synneg'), ax=axes[1][1])
venn3(metfrag, ('hilneg', 'hilpos', 'synneg'), ax=axes[1][2])
axes[0][0].set_title('All')
axes[1][0].set_title('Metlin')
axes[1][1].set_title('HMDB')
axes[1][2].set_title('MetFrag')
plt.show()
def plot_hil_neg():
observation_file = hil_neg
metlin = read.metlin(observation_file)
hmdb = read.hmdb(observation_file)
metfrag = set(read.metfrag_with_scores(observation_file).keys())
metfrag_no_zeros = set(
read.metfrag_with_scores(observation_file,
keep_zero_scores=False).keys())
feature_compounds = set([f for f in features if str.startswith(f, 'C')])
"""
PLOTTING
"""
figure, axes = plt.subplots(2, 3)
venn3([features, metlin, hmdb], ('features', 'metlin', 'hmdb'),
ax=axes[0][0])
venn3([features, cofactors, metlin | hmdb],
('features', 'cofactors', 'observed'),
ax=axes[0][1])
venn3([cofactors, metlin, hmdb], ('cofactors', 'metlin', 'hmdb'),
ax=axes[0][2])
venn3([metfrag_no_zeros, cofactors, features],
('metfrag_no_zeros', 'cofactors', 'features'),
ax=axes[1][0])
venn3([
metfrag - cofactors, (metlin | hmdb) - cofactors, features - cofactors
], ('metfrag_with_zeros', 'observed', 'features'),
ax=axes[1][1])
venn3([
metfrag_no_zeros - cofactors,
(metlin | hmdb) - cofactors, features - cofactors
], ('metfrag', 'observed', 'features'),
ax=axes[1][2])
axes[0][0].set_title('hard_evidence')
axes[0][1].set_title('hard evidence cofactor analysis vs features')
axes[0][2].set_title('hard evidence cofactor analysis')
axes[1][0].set_title('metfrag cofactor analysis')
axes[1][1].set_title('metfrag with zero scores (exclude cofactors)')
axes[1][2].set_title('metfrag without zero scores (exclude cofactors)')
plt.show()
def summarize(pathway):
path_dict = read.get_model(data_path + 'model2.csv')
pathways = path_dict.keys()
features = read.get_metabolites(path_dict)
cofactors = read.get_cofactors(data_path + 'cofactors')
paths = (hil_neg, hil_pos, syn_neg)
metlin = []
hmdb = []
metfrag = []
for path in paths:
metlin.append(read.metlin(path))
hmdb.append(read.hmdb(path))
metfrag.append(
set(read.metfrag_with_scores(path, keep_zero_scores=False).keys()))
# remove non-features & cofactors
for i in range(len(paths)):
metlin[i] = (metlin[i] & features) - cofactors
hmdb[i] = (hmdb[i] & features) - cofactors
metfrag[i] = (metfrag[i] & features) - cofactors
all = []
for db in [metlin, hmdb, metfrag]:
all.append(reduce(lambda x, y: x | y, db))
def stats(pathway, metlin, hmdb, metfrag=set()):
cge = path_dict[pathway]
observed = ((metlin | hmdb | metfrag) & features) - cofactors
reverse_path_dict = read.reverse_dict(path_dict)
for compound in observed & cge:
print compound, len(reverse_path_dict[compound])
compounds = path_dict[pathway]
print "Prob(C05381)", read.metfrag_with_scores(list(paths))['C05381']
print pathway, len(compounds - cofactors)
print "Unobserved:"
stats(pathway, compounds - cofactors, set())
print "Full Stats:"
stats(pathway, all[0], all[1], all[2])
print "Hard Stats:"
stats(pathway, all[0], all[1], set())
def get_metabolite_sets():
""" Gets the sets of observed metabolites """
path_dict = read.get_model(data_path + 'model2.csv')
pathways = path_dict.keys()
features = read.get_metabolites(path_dict)
cofactors = read.get_cofactors(data_path + 'cofactors')
paths = (hil_neg, hil_pos, syn_neg)
metlin = []
hmdb = []
metfrag = []
for path in paths:
metlin.append(read.metlin(path))
hmdb.append(read.hmdb(path))
metfrag.append(
set(read.metfrag_with_scores(path, keep_zero_scores=False).keys()))
# remove non-features & cofactors
for i in range(len(paths)):
metlin[i] = (metlin[i] & features) - cofactors
hmdb[i] = (hmdb[i] & features) - cofactors
metfrag[i] = (metfrag[i] & features) - cofactors
all = []
for db in [metlin, hmdb, metfrag]:
all.append(reduce(lambda x, y: x | y, db))
return all
data_path = '../data/'
observation_file = data_path + 'HilNeg 0324 -- Data.csv'
cofactors = read.get_cofactors(data_path + 'cofactors')
path_dict = read.get_model(data_path + 'model2.csv', cofactors=cofactors)
pathways = path_dict.keys()
features = read.get_metabolites(path_dict)
evidence = read.metlin(observation_file)
evidence |= read.hmdb(observation_file)
evidence -= cofactors
features -= cofactors
evidence &= features
reverse_path_dict = read.reverse_dict(path_dict)
metfrag = read.metfrag(observation_file)
metfrag_evidence = read.dict_of_set(
read.metfrag_with_scores(observation_file, keep_zero_scores=False),
metfrag & features - cofactors - evidence)
evidence = {e: 1 for e in evidence}
rate_prior = 0.5
ap = {p: Gamma('p_' + p, rate_prior, 1) for p in pathways}
bmp = {
p: {
feat: Gamma('b_{' + p + ',' + feat + '}', ap[p], 1)
for feat in path_dict[p]
}
for p in pathways
}
y_bmp = {}
virtual = {}