def make_row_scoring(self):
"""returns the row scoring function"""
row_scoring = microarray.RowScoringFunction(
self.membership(), self.ratio_matrix,
lambda iteration: ROW_WEIGHT,
config_params=self.config_params)
# motifing
sequence_filters = []
background_file_prom = meme.global_background_file(
self.organism(), self.ratio_matrix.row_names(), 'upstream',
use_revcomp=True)
background_file_p3utr = meme.global_background_file(
self.organism(), self.ratio_matrix.row_names(), 'p3utr',
use_revcomp=True)
meme_suite_prom = meme.MemeSuite430(
max_width=MAX_MOTIF_WIDTH,
background_file=background_file_prom)
meme_suite_p3utr = meme.MemeSuite430(
max_width=MAX_MOTIF_WIDTH,
background_file=background_file_p3utr)
motif_scoring = motif.MemeScoringFunction(
self.organism(),
self.membership(),
self.ratio_matrix,
meme_suite_prom,
seqtype='upstream',
sequence_filters=sequence_filters,
scaling_func=lambda iteration: 0.0,
run_in_iteration=scoring.default_motif_iterations,
config_params=self.config_params)
weeder_scoring = motif.WeederScoringFunction(
self.organism(), self.membership(), self.ratio_matrix,
meme_suite_p3utr, 'p3utr',
scaling_func=lambda iteration: 0.0,
run_in_iteration=scoring.default_motif_iterations,
config_params=self.config_params)
motif_combiner = scoring.ScoringFunctionCombiner(
self.membership(), [motif_scoring, weeder_scoring],
scaling_func=lambda iteration: 0.5,
config_params=self.config_params)
network_scoring = nw.ScoringFunction(self.organism(),
self.membership(),
self.ratio_matrix,
lambda iteration: 0.0,
scoring.default_network_iterations,
config_params=self.config_params)
return scoring.ScoringFunctionCombiner(
self.membership(), [row_scoring, network_scoring, motif_combiner],
config_params=self.config_params)
def __setup_meme_suite(self, config_params):
background_file = None
bgmodel = None
meme_version = config_params['MEME']['version']
search_distance = config_params['search_distances'][self.seqtype]
if config_params['MEME']['global_background'] == 'True':
background_file, bgmodel = meme.global_background_file(
self.organism, self.ratios.row_names, self.seqtype,
bgorder=int(self.config_params['MEME']['background_order']))
# store background in results database
conn = sqlite3.connect(config_params['out_database'], 15, isolation_level='DEFERRED')
for order in bgmodel:
for subseq, pvalue in order.items():
conn.execute('insert into global_background (subsequence, pvalue) values (?,?)',
(subseq, pvalue))
conn.commit()
conn.close()
if meme_version == '4.3.0':
self.meme_suite = meme.MemeSuite430(self.config_params,
background_file=background_file,
bgmodel=bgmodel)
elif meme_version and (
meme_version.startswith('4.8') or meme_version.startswith('4.9') or
meme_version.startswith('4.10')):
self.meme_suite = meme.MemeSuite481(self.config_params,
background_file=background_file,
bgmodel=bgmodel)
else:
logging.error("MEME version %s currently not supported !", meme_version)
raise Exception("unsupported MEME version: '%s'" % meme_version)
self.__sequence_filters = [unique_filter, get_remove_low_complexity_filter(self.meme_suite),
get_remove_atgs_filter(search_distance)]
def __setup_meme_suite(self, config_params):
background_file = None
bgmodel = None
meme_version = config_params['MEME']['version']
search_distance = config_params['search_distances'][self.seqtype]
if config_params['MEME']['global_background'] == 'True':
background_file, bgmodel = meme.global_background_file(
self.organism,
self.ratios.row_names,
self.seqtype,
bgorder=int(self.config_params['MEME']['background_order']))
# store background in results database
conn = sqlite3.connect(config_params['out_database'],
15,
isolation_level='DEFERRED')
for order in bgmodel:
for subseq, pvalue in order.items():
conn.execute(
'insert into global_background (subsequence, pvalue) values (?,?)',
(subseq, pvalue))
conn.commit()
conn.close()
if meme_version == '4.3.0':
self.meme_suite = meme.MemeSuite430(
self.config_params,
background_file=background_file,
bgmodel=bgmodel)
elif meme_version and (meme_version.startswith('4.8')
or meme_version.startswith('4.9')
or meme_version.startswith('4.10')):
self.meme_suite = meme.MemeSuite481(
self.config_params,
background_file=background_file,
bgmodel=bgmodel)
else:
logging.error("MEME version %s currently not supported !",
meme_version)
raise Exception("unsupported MEME version: '%s'" % meme_version)
self.__sequence_filters = [
unique_filter,
get_remove_low_complexity_filter(self.meme_suite),
get_remove_atgs_filter(search_distance)
]
def make_row_scoring(self):
"""returns the row scoring function"""
row_scoring = microarray.RowScoringFunction(
self.membership(), self.ratio_matrix,
lambda iteration: ROW_WEIGHT,
config_params=self.config_params)
sequence_filters = []
background_file_prom = meme.global_background_file(
self.organism(), self.ratio_matrix.row_names(), 'promoter',
use_revcomp=True)
background_file_p3utr = meme.global_background_file(
self.organism(), self.ratio_matrix.row_names(), 'p3utr',
use_revcomp=True)
meme_suite_prom = meme.MemeSuite430(
max_width=MAX_MOTIF_WIDTH,
background_file=background_file_prom)
meme_suite_p3utr = meme.MemeSuite430(
max_width=MAX_MOTIF_WIDTH,
background_file=background_file_p3utr)
motif_scoring = motif.MemeScoringFunction(
self.organism(),
self.membership(),
self.ratio_matrix,
meme_suite_prom,
seqtype='promoter',
sequence_filters=sequence_filters,
pvalue_filter=motif.MinPValueFilter(-20.0),
scaling_func=lambda iteration: 0.0,
run_in_iteration=scoring.default_motif_iterations,
config_params=self.config_params)
network_scoring = nw.ScoringFunction(self.organism(),
self.membership(),
self.ratio_matrix,
lambda iteration: 0.0,
scoring.default_network_iterations,
config_params=self.config_params)
weeder_scoring = motif.WeederScoringFunction(
self.organism(), self.membership(), self.ratio_matrix,
meme_suite_p3utr, 'p3utr',
pvalue_filter=motif.MinPValueFilter(-20.0),
scaling_func=lambda iteration: 0.0,
run_in_iteration=scoring.default_motif_iterations,
config_params=self.config_params)
pita = se.SetType.read_csv('pita', 'human_data/pita_miRNA_sets.csv')
target_scan = se.SetType.read_csv(
'target_scan', 'human_data/targetScan_miRNA_sets.csv')
set_types = [pita, target_scan]
set_enrichment_scoring = se.ScoringFunction(
self.membership(),
self.ratio_matrix,
set_types,
lambda iteration: 0.0, 7,
config_params=self.config_params)
motif_combiner = scoring.ScoringFunctionCombiner(
self.membership(),
[motif_scoring, weeder_scoring],
scaling_func=lambda iteration: 0.5)
return scoring.ScoringFunctionCombiner(
self.membership(),
[row_scoring, motif_combiner, network_scoring,
set_enrichment_scoring])
def make_row_scoring(self):
"""makes a row scoring function on demand"""
# Default row scoring functions
row_scaling_fun = scoring.get_scaling(self, 'row_')
row_scoring = microarray.RowScoringFunction(
self.organism(),
self.membership(),
self.ratio_matrix,
scaling_func=row_scaling_fun,
schedule=self["row_schedule"],
config_params=self.config_params)
row_scoring_functions = [row_scoring]
if self['domotifs']:
background_file = None
if self['global_background']:
background_file = meme.global_background_file(
self.organism(), self.ratio_matrix.row_names,
self['sequence_types'][0])
if self['meme_version'] == '4.3.0':
meme_suite = meme.MemeSuite430(background_file=background_file)
elif (self['meme_version']
and (self['meme_version'].startswith('4.8')
or self['meme_version'].startswith('4.9'))):
meme_suite = meme.MemeSuite481(background_file=background_file)
else:
logging.error("MEME version %s currently not supported !",
self['meme_version'])
raise Exception("unsupported MEME version")
sequence_filters = [
motif.unique_filter,
motif.get_remove_low_complexity_filter(meme_suite),
motif.get_remove_atgs_filter(
self['search_distances']['upstream'])
]
motif_scaling_fun = scoring.get_scaling(self, 'motif_')
nmotif_fun = motif.num_meme_motif_fun(self)
motif_scoring = motif.MemeScoringFunction(
self.organism(),
self.membership(),
self.ratio_matrix,
meme_suite,
sequence_filters=sequence_filters,
scaling_func=motif_scaling_fun,
num_motif_func=nmotif_fun,
update_in_iteration=self['motif_schedule'],
motif_in_iteration=self['meme_schedule'],
config_params=self.config_params)
row_scoring_functions.append(motif_scoring)
if self['donetworks']:
network_scaling_fun = scoring.get_scaling(self, 'network_')
network_scoring = nw.ScoringFunction(
self.organism(),
self.membership(),
self.ratio_matrix,
scaling_func=network_scaling_fun,
schedule=self['network_schedule'],
config_params=self.config_params)
row_scoring_functions.append(network_scoring)
return scoring.ScoringFunctionCombiner(
self.organism(),
self.membership(),
row_scoring_functions,
config_params=self.config_params)
def make_row_scoring(self):
"""returns the row scoring function"""
row_scoring = microarray.RowScoringFunction(
self.membership(), self.ratio_matrix,
lambda iteration: ROW_WEIGHT,
config_params=self.config_params)
sequence_filters = []
background_file_upstream = meme.global_background_file(
self.organism(), self.ratio_matrix.row_names(), 'upstream',
use_revcomp=True)
background_file_downstream = meme.global_background_file(
self.organism(), self.ratio_matrix.row_names(), 'downstream',
use_revcomp=True)
meme_suite_upstream = meme.MemeSuite430(
max_width=MAX_MOTIF_WIDTH,
background_file=background_file_upstream)
meme_suite_downstream = meme.MemeSuite430(
max_width=MAX_MOTIF_WIDTH,
background_file=background_file_downstream)
upstream_motif_scoring = motif.MemeScoringFunction(
self.organism(),
self.membership(),
self.ratio_matrix,
meme_suite_upstream,
seqtype='upstream',
sequence_filters=sequence_filters,
pvalue_filter=motif.MinPValueFilter(-20.0),
scaling_func=lambda iteration: 0.0,
run_in_iteration=motif_iterations,
config_params=self.config_params)
downstream_motif_scoring = motif.MemeScoringFunction(
self.organism(),
self.membership(),
self.ratio_matrix,
meme_suite_downstream,
seqtype='downstream',
sequence_filters=sequence_filters,
pvalue_filter=motif.MinPValueFilter(-20.0),
scaling_func=lambda iteration: 0.0,
run_in_iteration=motif_iterations,
config_params=self.config_params)
# NOTE: it looks like Weeder scoring isn't fully implemented yet? At this time, all other run configs seem to be passing MemeSuite430 into the WeederScoringFuncion
# weeder_scoring = motif.WeederScoringFunction(
# self.organism(), self.membership(), self.ratio_matrix,
# meme_suite_downstream, 'downstream',
# pvalue_filter=motif.MinPValueFilter(-20.0),
# scaling_func=lambda iteration: 0.0,
# run_in_iteration=motif_iterations,
# config_params=self.config_params)
network_scoring = nw.ScoringFunction(self.organism(),
self.membership(),
self.ratio_matrix,
network_iterations,
scoring.default_network_iterations,
config_params=self.config_params)
motif_combiner = scoring.ScoringFunctionCombiner(
self.membership(),
[upstream_motif_scoring, downstream_motif_scoring],
scaling_func=lambda iteration: 0.5)
return scoring.ScoringFunctionCombiner(
self.membership(),
[row_scoring, motif_combiner, network_scoring])
def meme_suite(self, seqtype):
"""upstream meme suite"""
background_file = meme.global_background_file(
self.organism(), self.ratio_matrix.row_names, seqtype, use_revcomp=True)
return meme.MemeSuite430(max_width=MAX_MOTIF_WIDTH,
background_file=background_file)
def make_row_scoring(self):
"""makes a row scoring function on demand"""
# Default row scoring functions
row_scaling_fun = scoring.get_scaling(self, 'row_')
row_scoring = microarray.RowScoringFunction(self.organism(),
self.membership(), self.ratio_matrix,
scaling_func=row_scaling_fun,
schedule=self["row_schedule"],
config_params=self.config_params)
row_scoring_functions = [row_scoring]
if self['domotifs']:
background_file = None
if self['global_background']:
background_file = meme.global_background_file(
self.organism(), self.ratio_matrix.row_names,
self['sequence_types'][0])
if self['meme_version'] == '4.3.0':
meme_suite = meme.MemeSuite430(background_file=background_file)
elif (self['meme_version'] and
(self['meme_version'].startswith('4.8') or
self['meme_version'].startswith('4.9'))):
meme_suite = meme.MemeSuite481(background_file=background_file)
else:
logging.error("MEME version %s currently not supported !", self['meme_version'])
raise Exception("unsupported MEME version")
sequence_filters = [
motif.unique_filter,
motif.get_remove_low_complexity_filter(meme_suite),
motif.get_remove_atgs_filter(self['search_distances']['upstream'])]
motif_scaling_fun = scoring.get_scaling(self, 'motif_')
nmotif_fun = motif.num_meme_motif_fun(self)
motif_scoring = motif.MemeScoringFunction(self.organism(),
self.membership(),
self.ratio_matrix,
meme_suite,
sequence_filters=sequence_filters,
scaling_func=motif_scaling_fun,
num_motif_func=nmotif_fun,
update_in_iteration=self['motif_schedule'],
motif_in_iteration=self['meme_schedule'],
config_params=self.config_params)
row_scoring_functions.append(motif_scoring)
if self['donetworks']:
network_scaling_fun = scoring.get_scaling(self, 'network_')
network_scoring = nw.ScoringFunction(self.organism(),
self.membership(),
self.ratio_matrix,
scaling_func=network_scaling_fun,
schedule=self['network_schedule'],
config_params=self.config_params)
row_scoring_functions.append(network_scoring)
return scoring.ScoringFunctionCombiner(
self.organism(),
self.membership(),
row_scoring_functions,
config_params=self.config_params)
