def main():
"""Run all sets of hypotheses experiments."""
args = get_args()
ntobj = cx.namedtuple("NtRunParams", "num_experiments num_sims dotsize")
#pylint: disable=bad-whitespace, no-member
lst_experiment_cnts = [
ntobj._make([500, 2500, {
'fdr_actual': 0.70,
'sensitivity': 0.40
}]), # 00:NN:NN
ntobj._make([500, 1000, {
'fdr_actual': 0.90,
'sensitivity': 0.50
}]), # 02:37:NN
ntobj._make([100, 1000, {
'fdr_actual': 1.20,
'sensitivity': 0.65
}]), # 00:24:31
ntobj._make([20, 200, {
'fdr_actual': 2.00,
'sensitivity': 2.00
}]), # 00:01:00
ntobj._make([10, 10, {
'fdr_actual': 2.00,
'sensitivity': 2.00
}]), # 00:00:02
]
ntd = lst_experiment_cnts[args['idx_experiment_cnts']]
# 0.01, 0.03, 0.05
max_sigpvals_super = [0.00001, 0.00001, 0.00001]
for perc_super in [0, 25, 50, 75]:
run(args, ntd, max_sigpvals_super, perc_super)
def main():
"""Run all sets of hypotheses experiments."""
args = get_args()
ntobj = cx.namedtuple("NtRunParams", "num_experiments num_sims dotsize")
#pylint: disable=bad-whitespace, no-member
lst_experiment_cnts = [
ntobj._make([500, 2500, {
'fdr_actual': 0.70,
'sensitivity': 0.40
}]), # 00:NN:NN
ntobj._make([500, 1000, {
'fdr_actual': 0.90,
'sensitivity': 0.50
}]), # 02:37:NN
ntobj._make([100, 1000, {
'fdr_actual': 1.20,
'sensitivity': 0.65
}]), # 00:24:31
ntobj._make([20, 200, {
'fdr_actual': 2.00,
'sensitivity': 2.00
}]), # 00:01:00
ntobj._make([10, 10, {
'fdr_actual': 2.00,
'sensitivity': 2.00
}]), # 00:00:02
]
ntd = lst_experiment_cnts[args['idx_experiment_cnts']]
lst_max_sigpvals = [
[0.01, 0.03, 0.05],
[0.0001, 0.001, 0.01],
]
max_sigpvals = lst_max_sigpvals[args['idx_max_sigpvals']]
methods = [ # Time for one ExperimentSet (100, 1000) H:MM:SS
#'bonferroni', # 0) Bonferroni one-step correction 0:00:09
#'sidak', # 1) Sidak one-step correction 0:00:09
#'holm-sidak', # 2) Holm-Sidak step-down method using Sidak adjustments 0:00:12
##holm', # 3) Holm step-down method using Bonferroni adjustments 2:29:59
#'simes-hochberg', # 4) Simes-Hochberg step-up method (independent) 0:00:11
'hommel', # 5) Hommel closed method based on Simes tests (non-negative)
'fdr_by', # 7) FDR Benjamini/Yekutieli (negative)
'fdr_tsbh', # 8) FDR 2-stage Benjamini-Hochberg (non-negative)
'fdr_tsbky', # 9) FDR 2-stage Benjamini-Krieger-Yekutieli (non-negative)
'fdr_gbs', # 10) FDR adaptive Gavrilov-Benjamini-Sarkar
'fdr_bh', # 6) FDR Benjamini/Hochberg (non-negative)
]
#main(args, ntd.num_experiments, ntd.num_sims, ntd.dotsize)
for method in methods:
run(args, ntd, method, max_sigpvals)
def main():
"""Simulate Gene Ontology Enrichment Analyses."""
args = get_args()
obj = Basename()
modules = [
'pkggosim.data.orig_noprune_enriched_ntn2_p0_100to000_004to124_N00002_00002_genes',
'pkggosim.data.orig_noprune_enriched_ntn2_p0_100to000_004to124_N00002_00002_goids',
]
# randomize_truenull_assc = args.get('randomize_truenull_assc', 'orig')
# study_bg = "humoral_rsp"
# popnullmaskout = ['immune', 'viral_bacteria']
# # Gene Ontology Data
# genes_mus = ensm2nt.keys() # Population genes
# params = {
# 'log' : None if ntd.num_experiments > 4 else sys.stdout,
# 'prefix' : 'fig_goea_{RND}'.format(RND=randomize_truenull_assc),
# 'randomize_truenull_assc' : randomize_truenull_assc,
# 'seed' : args.get('randomseed', None),
# 'alpha' : 0.05,
# 'method' : 'fdr_bh',
# 'propagate_counts' : args.get('propagate_counts', False),
# 'genes_population':genes_mus,
# 'genes_study_bg':import_genes(study_bg),
# 'goids_study_bg':import_goids(study_bg),
# 'genes_popnullmaskout':import_genes_all(popnullmaskout),
# 'association_file':'gene_association.mgi',
# 'perc_nulls' : [100, 75, 50, 25, 0],
# #'num_genes_list' : [4, 16, 64, 128],
# #'num_genes_list' : [4, 8, 16, 24, 32, 40, 48, 56, 64],
num_genes_list = args['genes']
# 'num_experiments' : ntd.num_experiments, # Num. of simulated FDR ratios per experiment set
# 'num_sims' : ntd.num_sims} # Number of sims per experiment; used to create one FDR ratio
# objparams = RunParams(params)
# #### obj = ExperimentsAll(objparams) # RunParams
# title_cur = objparams.get_title()
# prt.write("TITLE: {S}\n".format(S=title_cur))
# prt.write("GENES: {S}\n".format(S=params['num_genes_list']))
# prt.write("randomize_truenull_assc: {S}\n".format(S=randomize_truenull_assc))
# prt.write("{NT}\n".format(NT=ntd))
#### rpt_items = ['fdr_actual', 'sensitivity', 'specificity', 'pos_pred_val', 'neg_pred_val']
plt_items = ['fdr_actual', 'sensitivity', 'specificity']
pltargs = {'dotsize':None, 'title':args['title'],
'dpi':600, 'img':'all',
'xlabel':'Number of Genes in a Study Group',
'ylabel':'Percentage of General Population Genes'}
for mod in modules:
obj.plt_mod('log/dat_goea_plot', mod, plt_items, pltargs)
def main():
"""Arguments for running all experiments."""
args = get_args()
nto = cx.namedtuple("NtRunParams", "num_experiments num_sims dotsize")
#pylint: disable=bad-whitespace, no-member, line-too-long
experiment_cnts = [
nto._make([
500, 1000, {
'fdr_actual': 0.70,
'sensitivity': 0.50,
'specificity': 0.50
}
]),
nto._make([
100, 1000, {
'fdr_actual': 0.95,
'sensitivity': 0.60,
'specificity': 0.60
}
]),
nto._make([
100, 30, {
'fdr_actual': 1.30,
'sensitivity': 0.60,
'specificity': 0.60
}
]),
nto._make([
50, 50, {
'fdr_actual': 2.00,
'sensitivity': 0.70,
'specificity': 0.70
}
]),
nto._make([
50, 20, {
'fdr_actual': 2.00,
'sensitivity': 1.00,
'specificity': 1.00
}
]), # 4:56
nto._make([
20, 20, {
'fdr_actual': 2.00,
'sensitivity': 2.00,
'specificity': 2.00
}
]), # 1:25
nto._make([
4, 4, {
'fdr_actual': 4.00,
'sensitivity': 3.00,
'specificity': 3.00
}
]), # 0:04 0:05
nto._make([
2, 2, {
'fdr_actual': 4.00,
'sensitivity': 3.00,
'specificity': 3.00
}
]), # 0:01 0:02
]
ntd = experiment_cnts[args['idx_experiment_cnts']]
run(args, ntd)
def main(run_all, prt=sys.stdout):
"""Simulate small set of GOEAs to see full TP/FP/FN/TN gene GO counts."""
args = get_args()
#pylint: disable=no-member
#ntd = nto._make([1, 2, {'fdr_actual':4.00, 'sensitivity':3.00, 'specificity':3.00}]) # 0:01
ntd = NTO._make(
[4, 4, {
'fdr_actual': 4.00,
'sensitivity': 3.00,
'specificity': 3.00
}])
randomize_truenull_assc = args.get('randomize_truenull_assc', 'orig')
fout_log = "log/plt_goea_small/plt_goea_small_{NAME}.log".format(
NAME=randomize_truenull_assc)
study_bg = "humoral_rsp"
popnullmaskout = ['immune', 'viral_bacteria']
# Gene Ontology Data
with open(os.path.join(REPO, fout_log), 'w') as log:
params = {
'img': 'pdf', # png or pdf
'title': args['title'],
'log': log,
'prefix': 'fig_goea_{RND}'.format(RND=randomize_truenull_assc),
'randomize_truenull_assc': randomize_truenull_assc,
'seed': args.get('randomseed', None),
'alpha': 0.05,
'method': 'fdr_bh',
'propagate_counts': args['propagate_counts'],
'genes_population': ensm2nt.keys(), # Mouse protein-coding genes
'genes_study_bg': import_genes(study_bg),
'goids_study_bg': import_goids(study_bg),
'genes_popnullmaskout': import_genes_all(popnullmaskout),
'association_file': 'gene_association.mgi',
'perc_nulls': [100, 75, 50, 25, 0] if run_all else [25],
'num_genes_list':
args['genes'], # [4, 16, 64, 128] if run_all else [128],
'num_experiments': ntd.
num_experiments, # Num. simulated FDR ratios per experiment set
'num_sims': ntd.num_sims
} # Number sims per experiment; used to create one FDR ratio
obj = ExperimentsAll(RunParams(params)) # RunParams
sys.stdout.write(" propagate_counts={P}\n".format(
P=obj.pobj.params['propagate_counts']))
title_cur = obj.pobj.get_title()
prt.write("TITLE: {S}\n".format(S=title_cur))
prt.write("GENES: {S}\n".format(S=params['num_genes_list']))
prt.write(
"randomize_truenull_assc: {S}\n".format(S=randomize_truenull_assc))
prt.write("{NT}\n".format(NT=ntd))
rpt_items = [
'fdr_actual', 'sensitivity', 'specificity', 'pos_pred_val',
'neg_pred_val'
]
plt_items = ['fdr_actual', 'sensitivity', 'specificity']
pltargs = {
'dotsize': ntd.dotsize,
'title': title_cur,
'xlabel': 'Number of Genes in a Study Group',
'ylabel': 'Percentage of General Population Genes',
#'ylim':{'fdr_actual':[-0.005, 0.50]}
}
obj.run_all(study_bg, rpt_items, plt_items, **pltargs)
obj.pobj.prt_summary(log)
obj.pobj.prt_summary(sys.stdout)
sys.stdout.write(" propagate_counts={P}\n".format(
P=obj.pobj.params['propagate_counts']))
sys.stdout.write(" WROTE: {LOG}\n".format(LOG=fout_log))
sys.stdout.write("ARGS: {ARGS}\n".format(ARGS=args))