def main(args):
if not len(args) == 2:
sys.exit("USAGE: python checkGeneList.py geneList.txt > outFile")
# TCGA hub
hub = "https://tcga.xenahubs.net"
# PanCan normalized dataset
dataset = "TCGA.KIRC.sampleMap/HiSeqV2_PANCAN"
# get sample IDs
samples = xena.dataset_samples(hub, dataset, None)
#read geneList into list
f = open(args[1])
line = f.readline()[:-1]
genes = []
while line != "":
gene = line
genes.append(gene)
line = f.readline()[:-1]
#print expression
nGenes = len(genes)
k = xena.dataset_gene_probe_avg(hub, dataset, samples, genes)
for i in range(0, nGenes):
if k[i]['position'] == []:
print k[i]['gene']
def main(args):
if not len(args) == 3:
sys.exit(
"USAGE: python getExpressionFromXena.py cancerType[KIRC] geneList.txt > outFile"
)
# TCGA hub
hub = "https://tcga.xenahubs.net"
# PanCan normalized dataset
dataset = "TCGA." + str(args[1]) + ".sampleMap/HiSeqV2_PANCAN"
# get sample IDs
samples = xena.dataset_samples(hub, dataset, None)
#read geneList into list
f = open(args[2])
line = f.readline()[:-1]
genes = []
while line != "":
gene = line
genes.append(gene)
line = f.readline()[:-1]
#print header
print 'Gene\t' + '\t'.join(samples)
#print expression
nGenes = len(genes)
k = xena.dataset_gene_probe_avg(hub, dataset, samples, genes)
for i in range(0, nGenes):
geneName = k[i]['gene']
scores = k[i]['scores'][0]
print geneName + '\t' + '\t'.join(str(x) for x in scores)
def download_data():
"""Download and preprocess the expression and survival data."""
# Transcriptomic data
hub = "https://gdc.xenahubs.net"
dataset = "TCGA-BRCA.htseq_fpkm-uq.tsv"
samples = xena.dataset_samples(hub, dataset, None)
df_expression = get_expression_data(hub, dataset, samples)
# Survival data
df_survival = get_survival_data(hub, dataset, samples)
# Phenotype data (unused for now) : could be used to add covariates
# df_phenotype = get_phenotype_data(hub, samples)
df_all = merge_data(df_expression, df_survival)
filtered_df = filter_outliers(df_all)
record_train_test(filtered_df)
return
def download_gdc_clinicals(
xena_hub: str,
dataset: str,
*,
rowKey: str = "SampleID",
headers: List = [
"sample_type.samples",
"days_to_death.demographic",
"days_to_last_follow_up.diagnoses",
"vital_status.demographic",
],
) -> pd.DataFrame:
"""
Clinical information contains many factor columns, where factor names are not
readily accessible. This function decodes factor numbers into category names.
Parameters
----------
xena_hub
Url of the data repository hub.
dataset
The dataset containing survival information on the repository hub.
rowKey
The column containing (unique) identifiers - typically the sample IDs.
headers
A list of column names that we want to retrieve.
Returns
-------
Dataframe with clinical information, including survival.
"""
samples = xena.dataset_samples(xena_hub, dataset, None)
pos, mat = xena.dataset_probe_values(xena_hub, dataset, samples, headers)
clinicals = dict()
clinicals[rowKey] = samples
for i in range(len(mat)):
clinicals[headers[i]] = mat[i]
clinicals = pd.DataFrame(clinicals)
clinicals = clinicals.set_index(rowKey)
return clinicals
def main(args):
if not len(args) == 3:
sys.exit(
"USAGE: python replaceGeneListWithXenaAliasNames.py cancerType[KIRC] geneList.txt > newGeneList.txt"
)
# TCGA hub
hub = "https://tcga.xenahubs.net"
# PanCan normalized dataset
dataset = "TCGA." + str(args[1]) + ".sampleMap/HiSeqV2_PANCAN"
# get sample IDs
samples = xena.dataset_samples(hub, dataset, None)
#read geneList into list
f = open(args[2])
line = f.readline()[:-1]
genes = []
while line != "":
gene = line
genes.append(gene)
line = f.readline()[:-1]
f.close()
#if in Xena, print gene. If not in Xena, determine if gene has an alias that is in Xena
nGenes = len(genes)
k = xena.dataset_gene_probe_avg(hub, dataset, samples, genes)
for i in range(0, nGenes):
if k[i]['position'] == [] or k[i]['scores'][0][0] == 'NaN':
# determine if gene has an alias in Xena
alias = findAliasInXena(hub, dataset, samples, k[i]['gene'])
if alias != '':
print alias
else:
print k[i]['gene']
def check_allsamples(
xena_hub: str,
dataset: str,
*,
n: Union[None, int] = None,
) -> List:
"""
Check what samples are available in the dataset.
Parameters
----------
xena_hub
Url of the data repository hub.
dataset
The dataset we want to check on.
n
Limit the number of samples returned. Set to None if all samples are needed.
Returns
-------
A list of all sample names.
"""
return xena.dataset_samples(xena_hub, dataset, n)
def get_fields_and_codes(host, dataset, samples, fields):
"get fields and resolve NA in the value"
return get_codes(host, dataset, fields, get_fields( host, dataset, samples, fields))
# dictionary with all hub links
xena.PUBLIC_HUBS
# pancanAtlas cohort
cohort = 'TCGA PanCanAtlas'
host = xena.PUBLIC_HUBS['pancanAtlasHub']
# get expression for GENES
expression_dataset = 'EB++AdjustPANCAN_IlluminaHiSeq_RNASeqV2.geneExp.xena'
samples = xena.dataset_samples(host, expression_dataset, None)
samples[0: 10]
expression = get_fields_and_codes(host,
expression_dataset,
samples,
GENES) # list of lists.
expression_by_gene = dict(zip(GENES, expression)) # index by gene.
[expression_by_gene.keys(), GENES[0], expression_by_gene[GENES[0]][0:10]]
# note that missing data is returned as 'NaN'. One might want to remap this to None or NaN, depending on the later analysis tools.
# get disease type and survival columns
survival_dataset = 'Survival_SupplementalTable_S1_20171025_xena_sp'
fields = ['cancer type abbreviation', 'OS', 'OS.time']
values = get_fields_and_codes(host,
survival_dataset,
def get_TCGA_surv(self): samples = xena.dataset_samples(self.host, self.surv_dataset, None) values = xena.dataset_fetch(self.host, self.surv_dataset, samples,['_EVENT','_TIME_TO_EVENT']) surv_df = pd.DataFrame(data=values, index=['Event', 'Time_to_event'], columns=samples).T return surv_df