def calc_enrichment(gene_list, GO_ID_list, total_unique_gene, GO_Term_list):
M = total_unique_gene
enriched_list = []
for term in GO_ID_list:
if len(GO_ID_list.get(term)) >= 20 and len(
GO_ID_list.get(term)) <= 2000:
pvalue, overlap = calc_pvalue(gene_list, GO_ID_list.get(term), M)
if len(overlap) > 1:
enriched_item = {
"go_id": term,
"name": GO_Term_list.get(term)[0],
"description": GO_Term_list.get(term)[1],
"pvalue": pvalue,
"overlap": overlap,
"genes_from_list": len(gene_list),
"genes_from_go": len(GO_ID_list.get(term))
}
enriched_list.append(enriched_item)
enriched_list.sort(key=lambda it: it['pvalue'])
for qvalue, it in itertools.izip(
fdr([it['pvalue'] for it in enriched_list], presorted=True),
enriched_list):
if math.fabs(qvalue) == 0:
it['qvalue'] = 1000
else:
it['qvalue'] = -math.log(qvalue, 10)
enriched_list.sort(key=lambda it: it['qvalue'], reverse=True)
return enriched_list
def gene_set_enrichment(gene_list, M=None):
'''
:param gene_list: list of gene symbols
:param M: total number of genes (derived from database, if None)
:return: filtered list of GO terms with p-value, q-value, and size of overlap
'''
client = pymongo.MongoClient()
if not M:
M = len(client.go.genes.distinct('gene'))
terms = list(client.go.genes.find({'gene': {'$in': list(gene_list)}}).distinct('go'))
terms = list(client.go.terms.find({'go': {'$in': terms}, 'n_genes': {'$gt': 2}}))
enriched = [dict(term.items() + zip(('pvalue', 'overlap'), calc_pvalue(gene_list, term['genes'], M))) for term in terms]
enriched.sort(key=lambda it: it['pvalue'])
for qvalue, it in itertools.izip(fdr([it['pvalue'] for it in enriched], presorted=True), enriched):
it['qvalue'] = qvalue
return enriched
def calc_enrichment(gene_list, GO_ID_list, total_unique_gene, GO_Term_list):
M = total_unique_gene
enriched_list = []
for term in GO_ID_list:
if len(GO_ID_list.get(term)) >= 20 and len(GO_ID_list.get(term)) <= 2000:
pvalue, overlap = calc_pvalue(gene_list, GO_ID_list.get(term), M)
if len(overlap) > 1:
enriched_item = {"go_id": term, "name":GO_Term_list.get(term)[0] ,"description":GO_Term_list.get(term)[1],
"pvalue": pvalue, "overlap": overlap, "genes_from_list": len(gene_list), "genes_from_go": len(GO_ID_list.get(term))}
enriched_list.append(enriched_item)
enriched_list.sort(key=lambda it: it['pvalue'])
for qvalue, it in itertools.izip(fdr([it['pvalue'] for it in enriched_list], presorted=True), enriched_list):
if math.fabs(qvalue) == 0:
it['qvalue'] = float("inf")
else:
it['qvalue'] = -math.log(qvalue, 10)
enriched_list.sort(key=lambda it: it['qvalue'], reverse=True)
return enriched_list
def gene_set_enrichment(gene_list, M=None):
'''
:param gene_list: list of gene symbols --> ENSG00000233636, ENSG00000129673, etc...
:param M: total number of genes (derived from database, if None)
:return: filtered list of set terms with p-value, q-value, and size of overlap
'''
client = pymongo.MongoClient()
if not M:
M = len(client.go.genes.distinct('gene'))
terms = list(
client.go.genes.find({
'gene': {
'$in': list(gene_list)
}
}).distinct('go')) #GO:0004059, GO:0006474, etc...
terms = list(
client.go.terms.find({
'go': {
'$in': terms
},
'n_genes': {
'$gt': 2
}
}))
enriched = [
dict(term.items() +
zip(('pvalue',
'overlap'), calc_pvalue(gene_list, term['genes'], M)))
for term in terms
]
enriched.sort(key=lambda it: it['pvalue'])
for qvalue, it in itertools.izip(
fdr([it['pvalue'] for it in enriched], presorted=True), enriched):
it['qvalue'] = qvalue
return enriched
strDirection = 'PWS < PFS'
#print('ROI1 = %s, ROI2 = %s: p = %f (%s)'%(b_rois[i1], b_rois[i2], p, strDirection))
#print('ROI1 = %s, ROI2 = %s: p = %f (%s)'%(roi1, roi2, p, strDirection))
t_showZScores['PFS'].append(z_PFS)
t_showZScores['PWS'].append(z_PWS)
t_sigROIs['roi1'].append(roi1)
t_sigROIs['roi2'].append(roi2)
k = k + 1
print('Between-group comparisons of correlation z-scores')
if bFDR:
signifThresh = fdr(all_ps, FDR)
print('**FDR** (q = %f) --> signifThresh = %f\n'%(FDR, signifThresh))
else:
print('Pre-specified signifThresh = %f\n'%signifThresh);
### Determine the significant differences ###
keep_idx = []
for i1 in range(nTotComps):
if all_ps[i1] < signifThresh:
keep_idx.append(i1)
nSigDiff = len(keep_idx)
for idx in keep_idx:
p = all_ps[idx]
zScores_PFS = t_showZScores['PFS'][idx]
zScores_PWS = t_showZScores['PWS'][idx]
