def load_data(self, data_paths, method=None):
columns = []
col_names = []
row_name_index = {}
for col_index, data_file in enumerate(data_paths):
self.log.debug(" > {0}".format(data_file))
names = []
values = []
with tsv.open(data_file, "r") as f:
col_name, ext = os.path.splitext(os.path.basename(data_file))
params = tsv.params(f)
if "slice" in params:
col_name = params["slice"]
if "method" in params:
if method is None:
method = params["method"]
elif method != params["method"]:
self.log.warn("Different method of computation used for file {0}".format(data_file))
for name, value in tsv.lines(f, (str, float), header=True, null_value="-"):
if len(name) == 0:
self.log.warn("Empty identifier detected")
continue
if name not in row_name_index:
row_name_index[name] = len(row_name_index)
names += [name]
values += [value]
col_names += [col_name]
columns += [(names, values)]
num_cols = len(columns)
num_rows = len(row_name_index)
row_names = [None] * num_rows
for name, index in row_name_index.items():
row_names[index] = name
data = np.empty((num_rows, num_cols))
data[:] = np.nan
for col_index, (names, values) in enumerate(columns):
for i, name in enumerate(names):
data[row_name_index[name], col_index] = values[i]
return row_names, col_names, data, method
def drivers():
log = task.logger
config = GlobalConfig(task.conf)
paths = PathsConfig(config)
db_path = paths.results_path("drivers.db")
db = SigDb(db_path)
db.open()
log.info("Variants ...")
path = paths.combination_path("recurrences", "variant_gene-global-all.tsv.gz")
with tsv.open(path, "r") as f:
types = (str, str, int, str)
for fields in tsv.lines(f, types, columns=("CHR", "STRAND", "START", "ALLELE"), header=True):
chr, strand, start, allele = fields[:4]
db.add_variant(chr, start)
log.info("Genes ...")
gene_sites = {}
gene_fm = set()
gene_clust = set()
#SPECIAL_THRESHOLD = ["C18", "C34"]
SPECIAL_THRESHOLD = []
log.info(" OncodriveFM ...")
filename_re = re.compile(r"gene-cancer_site-(.+)\.tsv.gz")
base_path = paths.combination_path("oncodrivefm")
for path in os.listdir(base_path):
m = filename_re.match(path)
if not m:
continue
cancer_site_code = m.group(1)
if cancer_site_code in SPECIAL_THRESHOLD:
threshold = 1e-6
else:
threshold = 0.01
with tsv.open(os.path.join(base_path, path), "r") as f:
params = tsv.params(f)
cancer_site_name = params["group_long_name"]
for fields in tsv.lines(f, (str, float), columns=("ID", "QVALUE"), header=True):
gene, qvalue = fields
if qvalue < threshold:
add_cancer_site(gene_sites, gene, cancer_site_code, cancer_site_name)
gene_fm.add(gene)
log.info(" OncodriveCLUST ...")
filename_re = re.compile(r"cancer_site-(.+)\.tsv.gz")
base_path = paths.combination_path("oncodriveclust")
for path in os.listdir(base_path):
m = filename_re.match(path)
if not m:
continue
cancer_site_code = m.group(1)
with tsv.open(os.path.join(base_path, path), "r") as f:
params = tsv.params(f)
cancer_site_name = params["group_long_name"]
for fields in tsv.lines(f, (str, float), columns=("ID", "QVALUE"), header=True):
gene, qvalue = fields
if qvalue < 0.05:
add_cancer_site(gene_sites, gene, cancer_site_code, cancer_site_name)
gene_clust.add(gene)
log.info(" Updating db ...")
sig_genes = gene_fm | gene_clust
for gene in sig_genes:
db.add_gene(gene, gene in gene_fm, gene in gene_clust)
log.info("Saving driver genes cancer sites dataset ...")
path = paths.results_path("gene-driver_cancer_sites.tsv")
log.debug("> {}".format(path))
with open(path, "w") as f:
tsv.write_param(f, "date", datetime.now())
tsv.write_line(f, "GENE_ID", "FM", "CLUST", "CANCER_SITES_COUNT", "CANCER_SITE_CODES", "CANCER_SITE_NAMES")
for gene, sites in gene_sites.items():
tsv.write_line(f, gene,
1 if gene in gene_fm else 0,
1 if gene in gene_clust else 0,
len(sites),
", ".join(sorted([code for code, name in sites])),
", ".join(sorted([name for code, name in sites])))
db.commit()
db.close()
