def parseGeo(filename):
"""parseGeo(filename): open a GEO soft file and return the records it contains."""
try:
fh = gzip.open(filename)
records = Geo.parse(fh)
return records
except:
try:
fh = open(filename)
records = Geo.parse(fh)
return records
except:
print "Could not open filename"
sys.exit(3)
def parseGeo(filename):
'''parseGeo(filename): open a GEO soft file and return the records it contains.'''
try:
fh = gzip.open(filename)
records = Geo.parse(fh)
return (records)
except:
try:
fh = open(filename)
records = Geo.parse(fh)
return (records)
except:
print 'Could not open filename'
sys.exit(3)
def parseGeo(filename):
'''parseGeo(filename): open a GEO soft file and return the records it contains.'''
try:
fh = gzip.open(filename)
records = Geo.parse(fh)
return(records)
except:
try:
fh = open(filename)
records = Geo.parse(fh)
return(records)
except:
print 'Could not open filename'
sys.exit(3)
def parseGeo(filename):
try:
with gzip.open(filename) as fh:
records = Geo.parse(fh)
for record in records:
print record
except:
try:
with open(filename) as fh:
records = Geo.parse(fh)
for record in records:
print record
except:
print 'Could not open filename'
sys.exit(3)
def parseGeo(filename):
try:
with gzip.open(filename) as fh:
records = Geo.parse(fh)
for record in records:
print record
except:
try:
with open(filename) as fh:
records = Geo.parse(fh)
for record in records:
print record
except:
print 'Could not open filename'
sys.exit(3)
def read_gene_expr_data(geneset_dict):
"""
Read gene expression data from file. Returns
X - array of unnormalized gene expression data, grouped by genesets
y - control and disease labels
geneset - filtered genesets that match the given geneset dictionary, same order as returned X array
"""
handle = open(GENE_EXPR_FILENAME)
records = Geo.parse(handle)
# gsm ids of the normal subjects
normal_subjects = []
# geneset row ids
X_groups = {}
for k in geneset_dict.keys():
X_groups[k] = set()
X = []
y = []
i = 0
for record in records:
i += 1
if i == 3:
# Read patient labels so we can make the y vector
attr = record.entity_attributes
assert (attr["subset_description"] == "normal")
normal_subjects = attr["subset_sample_id"].split(",")
if i == 7:
# Read actual gene expression data
col_names = record.table_rows[0]
gsm_idxs = []
for idx, col_name in enumerate(col_names):
if "GSM" == col_name[0:3]:
gsm_idxs.append(idx)
# populate the y matrix
# 1 means diseased. 0 means control.
y.append(CONTROL_LABEL if col_name in
normal_subjects else DISEASE_LABEL)
geneid_idx = col_names.index("Gene ID")
feature_idx = 0
for row in record.table_rows[1:]:
geneid = row[geneid_idx]
geneset = get_geneset_from_dict(geneset_dict, geneid)
if geneset is not None:
# add feature idx to correct geneset
X_groups[geneset].add(feature_idx)
# append the gene expression data
X.append([float(row[i]) for i in gsm_idxs])
feature_idx += 1
# Make feature groups
X = np.matrix(X).T
X_genesets = []
genesets_included = []
for geneset_key, geneset_col_idxs in X_groups.iteritems():
if len(geneset_col_idxs) == 0:
continue
X_genesets.append(X[:, list(geneset_col_idxs)])
genesets_included.append(geneset_key)
y = np.matrix(y).T
return X_genesets, y, genesets_included
from Bio import Geo
handle = open('./GDS3292/GDS3292_full.soft')
records = Geo.parse(handle)
for record in records:
print record
handle.close()
def read_gene_expr_data(geneset_dict):
"""
Read gene expression data from file. Returns
X - array of unnormalized gene expression data, grouped by genesets
y - control and disease labels
geneset - filtered genesets that match the given geneset dictionary, same order as returned X array
"""
handle = open(GENE_EXPR_FILENAME)
records = Geo.parse(handle)
# gsm ids of the normal subjects
normal_subjects = []
# geneset row ids
X_groups = {}
for k in geneset_dict.keys():
X_groups[k] = set()
X = []
y = []
i = 0
for record in records:
i += 1
if i == 3:
# Read patient labels so we can make the y vector
attr = record.entity_attributes
assert(attr["subset_description"] == "normal")
normal_subjects = attr["subset_sample_id"].split(",")
if i == 7:
# Read actual gene expression data
col_names = record.table_rows[0]
gsm_idxs = []
for idx, col_name in enumerate(col_names):
if "GSM" == col_name[0:3]:
gsm_idxs.append(idx)
# populate the y matrix
# 1 means diseased. 0 means control.
y.append(CONTROL_LABEL if col_name in normal_subjects else DISEASE_LABEL)
geneid_idx = col_names.index("Gene ID")
feature_idx = 0
for row in record.table_rows[1:]:
geneid = row[geneid_idx]
geneset = get_geneset_from_dict(geneset_dict, geneid)
if geneset is not None:
# add feature idx to correct geneset
X_groups[geneset].add(feature_idx)
# append the gene expression data
X.append([float(row[i]) for i in gsm_idxs])
feature_idx += 1
# Make feature groups
X = np.matrix(X).T
X_genesets = []
genesets_included = []
for geneset_key, geneset_col_idxs in X_groups.iteritems():
if len(geneset_col_idxs) == 0:
continue
X_genesets.append(X[:, list(geneset_col_idxs)])
genesets_included.append(geneset_key)
y = np.matrix(y).T
return X_genesets, y, genesets_included