def vec_enrich(vec, gene_ids, quantile, gene_sets):
ind = np.quantile(vec, quantile) > vec
print("... {} features selected...".format(sum(ind)))
genes = gene_ids[ind]
# remove ens id version
genes = [re.sub("\\..*$", "", g) for g in genes]
print("Mapping to gene names...")
# map ens ids to gene symbols
bm = Biomart()
bm_result = bm.query(
dataset="hsapiens_gene_ensembl",
attributes=[
"ensembl_gene_id",
"external_gene_name",
"entrezgene_id",
"go_id",
],
filters={"ensembl_gene_id": genes},
)
gene_symbols = list(bm_result["external_gene_name"].unique())
print("Calculating enrichment...")
enr = gp.enrichr(
gene_list=gene_symbols,
gene_sets=gene_sets,
organism="Human",
cutoff=0.05,
)
return enr
def get_background(self):
"""get background gene"""
# input is a file
if os.path.isfile(self.background):
with open(self.background) as b:
bg2 = b.readlines()
bg = [g.strip() for g in bg2]
return set(bg)
# package included data
DB_FILE = resource_filename("gseapy", "data/{}.background.genes.txt".format(self.background))
if os.path.exists(DB_FILE):
df = pd.read_csv(DB_FILE,sep="\t")
else:
# background is a biomart database name
self._logger.warning("Downloading %s for the first time. It might take a couple of miniutes."%self.background)
bm = Biomart()
df = bm.query(dataset=self.background)
df.dropna(subset=['go_id'], inplace=True)
self._logger.info("Using all annotated genes with GO_ID as background: %s"%self.background)
df.dropna(subset=['entrezgene_id'], inplace=True)
# input id type: entrez or gene_name
if self._isezid:
bg = df['entrezgene_id'].astype(int)
else:
bg = df['external_gene_name']
return set(bg)
def get_background(self):
"""get background gene"""
DB_FILE = resource_filename("gseapy", "data/{}.background.genes.txt".format(self.background))
filename = os.path.join(DEFAULT_CACHE_PATH, "{}.background.genes.txt".format(self.background))
if os.path.exists(filename):
df = pd.read_table(filename)
elif os.path.exists(DB_FILE):
df = pd.read_table(DB_FILE)
else:
self._logger.warning("Downloading %s for the first time. It might take a couple of miniutes."%self.background)
bm = Biomart()
df = bm.query(dataset=self.background)
df.dropna(subset=['go_id'], inplace=True)
self._logger.info("using all annotated genes with GO_ID as background genes")
df.dropna(subset=['entrezgene'], inplace=True)
return df
adata = sc.AnnData(adata)
adata.obs["group"]=group.astype(np.str)
# rank gene by importance for clusters
sc.tl.rank_genes_groups(adata, "group", n_genes = 500)
r = adata.uns['rank_genes_groups']['names']
#pd.DataFrame.from_records(r).to_csv('marker_genes_group.csv')
#", ".join(pd.DataFrame.from_records(r)['1'].values)
for x in pd.DataFrame.from_records(r).columns:
print("group :"+x, end = "\r")
# rank gene by importance for clusters
glist = pd.DataFrame.from_records(r)[x].tolist()
bm = Biomart()
if not os.path.exists("test"):
os.makedirs("test")
results = bm.query(dataset='hsapiens_gene_ensembl',
attributes=['external_gene_name', 'go_id'],
filters={'hgnc_symbol': glist},
# save output file
filename="test/query_"+x+".results.txt")
enr = gp.enrichr(gene_list=glist,
description='test_name',
gene_sets=['KEGG_2016'],
outdir="test/enrichr_kegg_group"+x,
cutoff=0.5 # test dataset, use lower value from range(0,1)
)
def main():
GO_name = "TEST"
args = parser.parse_args()
input = args.input
# "score_D.txt"
p_threshold = float(args.threshold)
# 0.05
num = int(args.number)
############
# get GENE_COMM and final_score_D
GENE_COMM = []
final_score_D = []
file_D = input
fp = open(file_D, "r")
line = fp.readline()
line = fp.readline()
while (line != ""):
words = line.split("\t")
GENE_COMM.append(words[0])
final_score_D.append(words[-1].strip())
line = fp.readline()
fp.close()
###########
# sort data and get top 500 genes
data = np.array([GENE_COMM, final_score_D])
idex = np.lexsort([data[1, :]])
data_sorted = data[:, idex[::-1]]
GENES_top = data_sorted[0, :num]
SCORES_top = data_sorted[1, :num]
################
## use BioMart and convert ensembl_id to gene_symbol
bm = Biomart(verbose=False, host="asia.ensembl.org")
marts = bm.get_marts()
datasets = bm.get_datasets(mart='ENSEMBL_MART_ENSEMBL')
attrs = bm.get_attributes(dataset='hsapiens_gene_ensembl')
filters = bm.get_filters(dataset='hsapiens_gene_ensembl')
results = bm.query(dataset='hsapiens_gene_ensembl',
attributes=["ensembl_gene_id", "hgnc_symbol"],
filters={'ensembl_gene_id': GENES_top.tolist()},
filename="query.results.txt")
gene_List = []
hgnc_symbol_list = results.hgnc_symbol.tolist()
for gene in hgnc_symbol_list:
if (isinstance(gene, str)):
gene_List.append(gene)
libs = [
"GO_Biological_Process_2018", "GO_Cellular_Component_2018",
"GO_Molecular_Function_2018"
]
gseapy.enrichr(gene_list=gene_List,
description=GO_name,
gene_sets=libs,
outdir=GO_name)
##########
# fig for GO enrichment analysis. the threshold of adjusted p-value is less than 0.05, calculate -lg(adjusted p-value) and use gene num
TERMS = []
P_VALUE = []
GENES_NUM = []
file_GO = "GO_reports.txt"
fp_w = open(file_GO, "w")
mark = 0
for lib in libs:
GO_file = GO_name + "/" + lib + "." + GO_name + ".enrichr.reports.txt"
fp = open(GO_file, "r")
line = fp.readline()
if (mark == 0):
fp_w.write(line)
mark = 1
line = fp.readline()
while (line != ""):
fp_w.write(line)
words = line.split("\t")
term = words[1]
p_value = float(words[4])
genes_num = len(words[4].split(";"))
if (p_value < p_threshold):
TERMS.append(term.split(" ")[-1].strip('(').strip(')'))
P_VALUE.append(p_value)
GENES_NUM.append(genes_num)
line = fp.readline()
fp.close()
fp_w.close()
########
GO_value = []
for value in P_VALUE:
GO_value.append(-math.log(value, 10))
##############
fig = plt.figure(figsize=(9, 6))
matplotlib.rcParams['font.sans-serif'] = ['SimHei']
matplotlib.rcParams['axes.unicode_minus'] = False
plt.barh(range(len(GO_value)),
GO_value,
height=0.7,
color='steelblue',
alpha=0.8)
plt.yticks(range(len(GO_value)), TERMS)
plt.xlim(0, 10)
plt.xlabel("-LgP")
plt.title("GO enrichment")
for x, y in enumerate(GO_value):
plt.text(y + 0.2, x - 0.1, '%s' % y)
plt.show()
fig.savefig("GO enrichment barh.png")
#######
fig = plt.figure(figsize=(9, 6))
cm = plt.cm.get_cmap('RdYlGn')
NUM = []
for num in GENES_NUM:
NUM.append(100 * num)
sc = plt.scatter(GO_value, TERMS, c=GO_value, vmin=0, s=NUM, cmap=cm)
plt.colorbar(sc)
plt.show()
fig.savefig("GO enrichment scatter.png")