def random_samples_like(transcriptional_programs):
'Sample sets of targets and factors randomly that are the same size as those in the programs.'
from rpy import r
target_samples = []
factor_samples = []
for tp in transcriptional_programs:
if tp.targets_go_analysis:
target_samples.append(r.sample(tp.genes, len(tp.tp_targets)))
if tp.factors_go_analysis:
factor_samples.append(r.sample(tp.factors, len(tp.tp_factors)))
return target_samples, factor_samples
def random_samples_like(transcriptional_programs):
'Sample sets of targets and factors randomly that are the same size as those in the programs.'
from rpy import r
target_samples = []
factor_samples = []
for tp in transcriptional_programs:
if tp.targets_go_analysis:
target_samples.append(r.sample(tp.genes, len(tp.tp_targets)))
if tp.factors_go_analysis:
factor_samples.append(r.sample(tp.factors, len(tp.tp_factors)))
return target_samples, factor_samples
def generate_best_p_values(sample_sizes, num_samples_at_each_size, universe, go_data):
from rpy import r
for size in sample_sizes:
yield size, [
log10(calculate_best_p_value(go_data, r.sample(universe, size)))
for i in xrange(num_samples_at_each_size)
]
def generate_best_p_values(sample_sizes, num_samples_at_each_size, universe,
go_data):
from rpy import r
for size in sample_sizes:
yield size, [
log10(calculate_best_p_value(go_data, r.sample(universe, size)))
for i in xrange(num_samples_at_each_size)
]
def qqplot_density(samples, density_x, density_y):
from rpy import r
# LAME: should do better quantile calculation. r.quantile() returns a
# hard-to-use dictionary unfortunately. r.approx() as per ?quantile, maybe.
percs = list(np.arange(1,100, .1))
data_perc=np.percentile(samples, percs)
real_perc=np.percentile(r.sample(density_x, len(samples)*10, prob=density_y, replace=True), percs)
r.plot(real_perc, data_perc, xlab='',ylab='',main='');
r.abline(a=0,b=1,col='blue')
def qqplot_density(samples, density_x, density_y):
from rpy import r
# LAME: should do better quantile calculation. r.quantile() returns a
# hard-to-use dictionary unfortunately. r.approx() as per ?quantile, maybe.
percs = list(np.arange(1, 100, .1))
data_perc = np.percentile(samples, percs)
real_perc = np.percentile(
r.sample(density_x, len(samples) * 10, prob=density_y, replace=True),
percs)
r.plot(real_perc, data_perc, xlab='', ylab='', main='')
r.abline(a=0, b=1, col='blue')
#r.par(mfrow=[2,2]) # Exhaustive computation to compare to slice sampler log_concs = np.arange(-50,20,.1) lls = np.array([ loglik(log_conc) for log_conc in log_concs ]) mle = log_concs[lls.argmax()] print "MLE", np.exp(mle) lps = np.array([ logpost(log_conc) for log_conc in log_concs ]) map = log_concs[lps.argmax()] print "MAP", np.exp(map) #r.plot(log_concs, lls, xlab='x',ylab='ll') #plot(log_concs, lls, main='loglik') #plot(log_concs, lps, main='unnorm logpost') # For comparisons real_s = r.sample(log_concs,len(histories)*2,prob=np.exp(lps),replace=True) # Slice sampler h = mcmc.slice_sample(logpost, 0.0, 1.0, niter=1000) h = np.array(h) # Replications of single slice-sample #for i in range(500): # init = histories[i][-1] # h = mcmc.slice_sample(logpost, init, 1.0, niter=200) # histories[i] += h
