def make_weeder_scoring(self, seqtype, meme_suite, sequence_filters):
motif_scaling_fun = scoring.get_default_motif_scaling(self['num_iterations'])
return motif.WeederScoringFunction(
self.organism(), self.membership(), self.ratio_matrix,
meme_suite,
seqtype=seqtype,
sequence_filters=sequence_filters,
scaling_func=motif_scaling_fun,
num_motif_func=motif.default_nmotif_fun,
update_in_iteration=scoring.schedule(MOTIF_START_ITERATION, MOTIF_UPDATE_INTERVAL),
motif_in_iteration=scoring.schedule(MOTIF_START_ITERATION, MOTIF_COMPUTE_INTERVAL),
config_params=self.config_params)
def make_row_scoring(self):
"""makes a row scoring function on demand"""
# Default row scoring functions
row_scoring = microarray.RowScoringFunction(
self.membership(), self.ratio_matrix,
scaling_func=lambda iteration: self['row_scaling'],
config_params=self.config_params)
self.row_scoring = row_scoring
meme_suite = meme.MemeSuite430()
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_default_motif_scaling(
self['num_iterations'])
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=motif.default_nmotif_fun,
update_in_iteration=scoring.schedule(601, 3),
motif_in_iteration=scoring.schedule(600, 100),
#update_in_iteration=scoring.schedule(100, 10),
#motif_in_iteration=scoring.schedule(100, 100),
config_params=self.config_params)
self.motif_scoring = motif_scoring
network_scaling_fun = scoring.get_default_network_scaling(
self['num_iterations'])
network_scoring = nw.ScoringFunction(
self.organism(),
self.membership(),
self.ratio_matrix,
scaling_func=network_scaling_fun,
run_in_iteration=scoring.schedule(1, 7),
config_params=self.config_params)
self.network_scoring = network_scoring
row_scoring_functions = [row_scoring, motif_scoring, network_scoring]
return scoring.ScoringFunctionCombiner(
self.membership(),
row_scoring_functions,
config_params=self.config_params,
log_subresults=True)
def make_network_scoring(self, scaling_fun):
return nw.ScoringFunction(self.organism(),
self.membership(),
self.ratio_matrix,
scaling_func=scaling_fun,
run_in_iteration=scoring.schedule(1, 7),
config_params=self.config_params)
def __init__(self, membership, matrix, scaling_func=None,
run_in_iteration=scoring.schedule(1, 2),
config_params=None):
"""Create scoring function instance"""
scoring.ScoringFunctionBase.__init__(self, membership,
matrix, scaling_func,
run_in_iteration,
config_params)
self.cache_result = True
self.run_log = scoring.RunLog("row_scoring", config_params)
def make_set_scoring(self, scaling_fun, use_set_types):
set_types = []
if 'pita' in use_set_types:
set_types.append(se.SetType.read_csv('pita', 'human_data/gbmTCGA/pita_miRNA_sets_geneSymbol.csv'))
if 'target_scan' in use_set_types:
set_types.append(se.SetType.read_csv('target_scan',
'human_data/gbmTCGA/targetScan_miRNA_sets_geneSymbol.csv'))
return se.ScoringFunction(self.membership(), self.ratio_matrix,
set_types, scaling_fun,
scoring.schedule(1, 7),
config_params=self.config_params)
def make_row_scoring(self):
"""returns the row scoring function"""
row_scoring = microarray.RowScoringFunction(
self.membership(), self.ratio_matrix,
scaling_func=lambda iteration: self['row_scaling'],
config_params=self.config_params)
background_file_prom = meme_FS.global_background_file_FS(
self.organism(), self.ratio_matrix.row_names(), 'Promoter',
alphabet, alphabet_replacement,
use_revcomp=True, resultsDir = "/".join([self['meme_dir'], 'backgroundFiles']))
meme_suite = meme_FS.MemeSuite481(background_file=background_file_prom,
remove_tempfiles=True,
resultsDir = self['meme_dir'])
"""
background_file_p3utr = meme_FS.global_background_file_FS(
self.organism(), self.ratio_matrix.row_names(), '3pUTR',
alphabet, alphabet_replacement,
use_revcomp=True)
meme_suite_p3utr = meme_FS.MemeSuite430(background_file=background_file_p3utr)
"""
sequence_filters = [
motif_FS.unique_filter,
motif_FS.get_remove_low_complexity_filter(meme_suite),
motif_FS.get_remove_atgs_filter(self['search_distances']['Promoter'])]
sequence_filters = [motif_FS.get_remove_atgs_filter(self['search_distances']['Promoter'])]
motif_scaling_fun = scoring.get_default_motif_scaling(self['num_iterations'])
motif_scoring = motif_FS.MemeScoringFunction(
self.organism(),
self.membership(),
self.ratio_matrix,
meme_suite,
seqtype='Promoter',
sequence_filters=sequence_filters,
pvalue_filter=motif_FS.MinPValueFilter(-9),
scaling_func=motif_scaling_fun,
update_in_iteration=scoring.schedule(11, 3),
motif_in_iteration=scoring.schedule(10, 100),
config_params=self.config_params)
"""
motif_scoring_3pUTR = motif_FS.MemeScoringFunction(
self.organism(),
self.membership(),
self.ratio_matrix,
meme_suite,
seqtype='3pUTR',
sequence_filters=sequence_filters,
pvalue_filter=motif_FS.MinPValueFilter(-9),
scaling_func=motif_scaling_fun,
run_in_iteration=scoring.schedule(100, 50),
config_params=self.config_params)
motif_combiner = scoring.ScoringFunctionCombiner(
self.membership(),
[motif_scoring, motif_scoring_3pUTR],
scaling_func=lambda iteration: 0.5)
"""
'''
weeder_scoring = motif.WeederScoringFunction(
self.organism(), self.membership(), self.ratio_matrix,
meme_suite_p3utr, '3pUTR',
pvalue_filter=motif.MinPValueFilter(-20.0),
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)
'''
network_scaling_fun = scoring.get_default_network_scaling(self['num_iterations'])
network_scoring = nw.ScoringFunction(self.organism(),
self.membership(),
self.ratio_matrix,
scaling_func=network_scaling_fun,
run_in_iteration=scoring.schedule(1, 7),
config_params=self.config_params)
row_scoring_functions = [row_scoring, motif_scoring, network_scoring]
#row_scoring_functions = [row_scoring]
return scoring.ScoringFunctionCombiner(self.membership(),
row_scoring_functions,
log_subresults=True)