def analyze_paired_scores_with_and_without_priors(n_best=10):
nFitter = Fitter(Sigslope())
yFitter = Fitter(Sigslope(priors_name), 'normal')
nFits = get_all_fits(data,nFitter,allow_new_computation=False)
yFits = get_all_fits(data,yFitter,allow_new_computation=False)
score_pairs = [(f1.LOO_score, f2.LOO_score) for f1,f2 in iterate_fits(nFits, yFits)]
nScores, yScores = zip(*score_pairs)
_, pval = scipy.stats.wilcoxon(nScores, yScores)
pval = pval/2 # one sided p-value
print '*** wilcoxon signed rank p-value (one sided) = {:.3g}'.format(pval)
# find examples of best improvements
diffs = [(f2.LOO_score-f1.LOO_score, f1.LOO_score, f2.LOO_score, g, r) for dsname,g,r,f1,f2 in iterate_fits(nFits, yFits, R2_threshold=-1, return_keys=True)]
diffs.sort(reverse=True)
print 'Gene/Regions for which priors produce best R2 improvement:'
for i,(delta,R2_without, R2_with, g,r) in enumerate(diffs[:10]):
print '{i}) {g}@{r}, delta-R2={delta:.3g}. R2_without={R2_without:.3g}, R2_with={R2_with:.3g}'.format(**locals())
def do_fits(data, fitter, k_of_n, add_correlations, correlations_k_of_n):
n_correlation_iterations = 4 if add_correlations else 0
print """
==============================================================================================
==============================================================================================
==== Computing Fits with {}
==============================================================================================
==============================================================================================
""".format(fitter)
fits = get_all_fits(data, fitter, k_of_n, n_correlation_iterations=n_correlation_iterations, correlations_k_of_n=correlations_k_of_n)
return fits
def do_fits(data, fitter, k_of_n, add_correlations, correlations_k_of_n):
n_correlation_iterations = 4 if add_correlations else 0
print """
==============================================================================================
==============================================================================================
==== Computing Fits with {}
==============================================================================================
==============================================================================================
""".format(fitter)
fits = get_all_fits(data, fitter, k_of_n, n_correlation_iterations=n_correlation_iterations, correlations_k_of_n=correlations_k_of_n)
return fits
def plot_theta_diff_scatter(show_title=False):
yFitter = Fitter(Sigslope(priors_name),'normal')
nFitter = Fitter(Sigslope())
yFits = get_all_fits(data,yFitter)
nFits = get_all_fits(data,nFitter)
pairs = [(nFit.LOO_score,yFit.LOO_score) for nFit,yFit in iterate_fits(nFits,yFits)]
diff_pairs = [(n,y-n) for n,y in pairs if n is not None and y is not None]
n,d = zip(*diff_pairs)
fig = plt.figure()
ax = fig.add_axes([0.15,0.12,0.8,0.8])
ax.scatter(n, d, alpha=0.5)
xlims = ax.get_xlim()
ax.plot(xlims,[0, 0],'k--')
ax.set_xlim(*xlims)
if show_title:
ax.title(r'Improvement from prior on $\theta$ vs. baseline $R^2$', fontsize=fontsize)
ax.set_xlabel(r'$R^2$(no priors)', fontsize=fontsize)
ax.set_ylabel(r'$R^2$($\theta$) - $R^2$(no priors)', fontsize=fontsize)
ax.tick_params(axis='both', labelsize=fontsize)
return fig
def analyze_variant(theta,sigma):
theta_priors = priors_name if theta else None
sigma_prior = 'normal' if sigma else None
shape = Sigslope(theta_priors)
fitter = Fitter(shape,sigma_prior)
fits = get_all_fits(data,fitter,allow_new_computation=False)
LOO_scores = [f.LOO_score for f in iterate_fits(fits) if f.LOO_score is not None]
mu,sem = bootstrap(LOO_scores, np.mean)
return Bunch(
theta = theta,
sigma = sigma,
LOO_scores = LOO_scores,
mu = mu,
sem = sem,
)
def plot_theta_diff_scatter(show_title=False):
yFitter = Fitter(Sigslope(priors_name), 'normal')
nFitter = Fitter(Sigslope())
yFits = get_all_fits(data, yFitter)
nFits = get_all_fits(data, nFitter)
pairs = [(nFit.LOO_score, yFit.LOO_score)
for nFit, yFit in iterate_fits(nFits, yFits)]
diff_pairs = [(n, y - n) for n, y in pairs
if n is not None and y is not None]
n, d = zip(*diff_pairs)
fig = plt.figure()
ax = fig.add_axes([0.15, 0.12, 0.8, 0.8])
ax.scatter(n, d, alpha=0.5)
xlims = ax.get_xlim()
ax.plot(xlims, [0, 0], 'k--')
ax.set_xlim(*xlims)
if show_title:
ax.title(r'Improvement from prior on $\theta$ vs. baseline $R^2$',
fontsize=fontsize)
ax.set_xlabel(r'$R^2$(no priors)', fontsize=fontsize)
ax.set_ylabel(r'$R^2$($\theta$) - $R^2$(no priors)', fontsize=fontsize)
ax.tick_params(axis='both', labelsize=fontsize)
return fig
def analyze_paired_scores_with_and_without_priors(n_best=10):
nFitter = Fitter(Sigslope())
yFitter = Fitter(Sigslope(priors_name), 'normal')
nFits = get_all_fits(data, nFitter, allow_new_computation=False)
yFits = get_all_fits(data, yFitter, allow_new_computation=False)
score_pairs = [(f1.LOO_score, f2.LOO_score)
for f1, f2 in iterate_fits(nFits, yFits)]
nScores, yScores = zip(*score_pairs)
_, pval = scipy.stats.wilcoxon(nScores, yScores)
pval = pval / 2 # one sided p-value
print '*** wilcoxon signed rank p-value (one sided) = {:.3g}'.format(pval)
# find examples of best improvements
diffs = [(f2.LOO_score - f1.LOO_score, f1.LOO_score, f2.LOO_score, g, r)
for dsname, g, r, f1, f2 in iterate_fits(
nFits, yFits, R2_threshold=-1, return_keys=True)]
diffs.sort(reverse=True)
print 'Gene/Regions for which priors produce best R2 improvement:'
for i, (delta, R2_without, R2_with, g, r) in enumerate(diffs[:10]):
print '{i}) {g}@{r}, delta-R2={delta:.3g}. R2_without={R2_without:.3g}, R2_with={R2_with:.3g}'.format(
**locals())
def analyze_variant(theta, sigma):
theta_priors = priors_name if theta else None
sigma_prior = 'normal' if sigma else None
shape = Sigslope(theta_priors)
fitter = Fitter(shape, sigma_prior)
fits = get_all_fits(data, fitter, allow_new_computation=False)
LOO_scores = [
f.LOO_score for f in iterate_fits(fits) if f.LOO_score is not None
]
mu, sem = bootstrap(LOO_scores, np.mean)
return Bunch(
theta=theta,
sigma=sigma,
LOO_scores=LOO_scores,
mu=mu,
sem=sem,
)
def save_fits_and_create_html(data, fitter, fits=None, basedir=None,
do_genes=True, do_series=True, do_hist=True, do_html=True, only_main_html=False,
k_of_n=None,
use_correlations=False, correlations=None,
show_change_distributions=False,
html_kw=None,
figure_kw=None):
if fits is None:
fits = get_all_fits(data,fitter,k_of_n)
if basedir is None:
basedir = join(results_dir(), fit_results_relative_path(data,fitter))
if use_correlations:
basedir = join(basedir,'with-correlations')
if html_kw is None:
html_kw = {}
if figure_kw is None:
figure_kw = {}
print 'Writing HTML under {}'.format(basedir)
ensure_dir(basedir)
gene_dir = 'gene-subplot'
series_dir = 'gene-region-fits'
correlations_dir = 'gene-correlations'
scores_dir = 'score_distributions'
if do_genes and not only_main_html: # relies on the sharding of the fits respecting gene boundaries
plot_and_save_all_genes(data, fitter, fits, join(basedir,gene_dir), show_change_distributions)
if do_series and not only_main_html:
plot_and_save_all_series(data, fitter, fits, join(basedir,series_dir), use_correlations, show_change_distributions, figure_kw)
if do_hist and k_of_n is None and not only_main_html:
create_score_distribution_html(fits, use_correlations, join(basedir,scores_dir))
if do_html and k_of_n is None:
link_to_correlation_plots = use_correlations and correlations is not None
if link_to_correlation_plots and not only_main_html:
plot_and_save_all_gene_correlations(data, correlations, join(basedir,correlations_dir))
dct_pathways = load_17_pathways_breakdown()
pathway_genes = set.union(*dct_pathways.values())
data_genes = set(data.gene_names)
missing = pathway_genes - data_genes
b_pathways = len(missing) < len(pathway_genes)/2 # simple heuristic to create pathways only if we have most of the genes (currently 61 genes are missing)
create_html(
data, fitter, fits, basedir, gene_dir, series_dir, scores_dir, correlations_dir=correlations_dir,
use_correlations=use_correlations, link_to_correlation_plots=link_to_correlation_plots,
b_pathways=b_pathways, **html_kw
)
def get_onset_times(data, fitter, R2_threshold, b_force=False):
filename = join(cache_dir(),fit_results_relative_path(data,fitter) + '.pkl')
if isfile(filename):
print 'Loading onset distribution from {}'.format(filename)
with open(filename) as f:
bin_edges, change_vals = pickle.load(f)
else:
print 'Computing...'
fits = get_all_fits(data, fitter)
thetas = [fit.theta for fit in iterate_fits(fits, R2_threshold=R2_threshold)]
stages = [stage.scaled(age_scaler) for stage in dev_stages]
low = min(stage.from_age for stage in stages)
high = max(stage.to_age for stage in stages)
bin_edges, change_vals = compute_change_distribution(fitter.shape, thetas, low, high, n_bins=50)
print 'Saving result to {}'.format(filename)
ensure_dir(dirname(filename))
with open(filename,'w') as f:
pickle.dump((bin_edges,change_vals),f)
return bin_edges, change_vals
return mu,fit.LOO_score
lst_mu_R2 = [get_onset_time(r) for r in regions]
onset_times, lst_R2 = zip(*lst_mu_R2)
r,pval = spearmanr(onset_times, range(len(regions)))
return r,pval,lst_R2
lst_pathways = [
'serotonin',
'dopamine',
]
for pathway in lst_pathways:
data = GeneData.load('both').restrict_pathway(pathway).scale_ages(age_scaler)
shape = Sigmoid(priors='sigmoid_wide')
fitter = Fitter(shape, sigma_prior='normal')
fits = get_all_fits(data, fitter, allow_new_computation=False)
# R2_threshold = 0.5 YYY problem - we might be using bad fits.
regions = ['OFC', 'M1C', 'S1C', 'IPC', 'V1C']
scores = []
for g in data.gene_names:
r,pval,lst_R2 = get_gene_correlation(fits,g,regions)
scores.append( (g,r,pval,lst_R2) )
fig = plot_correlation_histogram(scores,pathway)
save_figure(fig,'{}/gradual-maturation-hist.png'.format(pathway,pathway), under_results=True, b_close=True)
for fR2 in [np.mean]: #[min,max,np.mean]:
fig = plot_scatter(scores, pathway, fR2)
save_figure(fig,'{}/gradual-maturation-scatter-{}.png'.format(pathway,fR2.__name__), under_results=True, b_close=True)
onset_times, lst_R2 = zip(*lst_mu_R2)
r, pval = spearmanr(onset_times, range(len(regions)))
return r, pval, lst_R2
lst_pathways = [
'serotonin',
'dopamine',
]
for pathway in lst_pathways:
data = GeneData.load('both').restrict_pathway(pathway).scale_ages(
age_scaler)
shape = Sigmoid(priors='sigmoid_wide')
fitter = Fitter(shape, sigma_prior='normal')
fits = get_all_fits(data, fitter, allow_new_computation=False)
# R2_threshold = 0.5 YYY problem - we might be using bad fits.
regions = ['OFC', 'M1C', 'S1C', 'IPC', 'V1C']
scores = []
for g in data.gene_names:
r, pval, lst_R2 = get_gene_correlation(fits, g, regions)
scores.append((g, r, pval, lst_R2))
fig = plot_correlation_histogram(scores, pathway)
save_figure(fig,
'{}/gradual-maturation-hist.png'.format(pathway, pathway),
under_results=True,
b_close=True)
ax.set_xlabel('z score', fontsize=fontsize)
ax.set_ylabel('probability', fontsize=fontsize)
ax.tick_params(axis='both', labelsize=fontsize)
return fig
cfg.verbosity = 1
age_scaler = LogScaler()
pathway = '17full'
data = GeneData.load('both').restrict_pathway(pathway).scale_ages(age_scaler)
data_shuffled = GeneData.load('both').restrict_pathway(pathway).scale_ages(
age_scaler).shuffle()
shape = Sigmoid('sigmoid_wide')
fitter = Fitter(shape, sigma_prior='normal')
fits = get_all_fits(data, fitter, allow_new_computation=False)
fits_shuffled = get_all_fits(data_shuffled,
fitter,
allow_new_computation=False)
R2_pairs = [(fit.LOO_score, fit2.LOO_score)
for fit, fit2 in iterate_fits(fits, fits_shuffled)]
R2 = np.array([r for r, r_shuffled in R2_pairs])
R2_shuffled = np.array([r_shuffled for r, r_shuffled in R2_pairs])
name = '{}-{}'.format(data.pathway, shape.cache_name())
fig = plot_score_distribution(R2, R2_shuffled)
save_figure(fig,
'RP/R2-distribution-{}.png'.format(name),
under_results=True,
b_close=True)
if __name__ == '__main__':
disable_all_warnings()
parser = get_common_parser()
parser.add_argument('--shape2', required=True, help='The shape to compare against', choices=allowed_shape_names())
parser.add_argument('--scaling2', help='The scaling used when fitting shape2. Default: none', choices=allowed_scaler_names())
parser.add_argument('--sigma_prior2', help='Prior to use for 1/sigma when fitting shape2. Default: None', choices=get_allowed_priors(is_sigma=True))
parser.add_argument('--priors2', help='The priors used for theta when fitting shape2. Default: None', choices=get_allowed_priors())
parser.add_argument('--filename', help='Where to save the figure. Default: results/comparison.png')
parser.add_argument('--show', help='Show figure and wait before exiting', action='store_true')
parser.add_argument('--ndiffs', type=int, default=5, help='Number of top diffs to show. Default=5.')
args = parser.parse_args()
data1, fitter1 = process_common_inputs(args)
data2 = get_data_from_args(args.dataset, args.pathway, args.from_age, args.scaling2, args.shuffle)
fitter2 = get_fitter_from_args(args.shape2, args.priors2, args.sigma_prior2)
fits1 = get_all_fits(data1,fitter1)
fits2 = get_all_fits(data2,fitter2)
print_diff_points(data1,fitter1,fits1, data2,fitter2,fits2, args.ndiffs)
fig = plot_comparison_scatter(data1,fitter1,fits1, data2,fitter2,fits2)
filename = args.filename
if filename is None:
ensure_dir(results_dir())
filename = join(results_dir(), 'shape_comparison.png')
save_figure(fig, filename)
if args.show:
plt.show(block=True)
#####################################################
GRs = [
('ABHD4','STC', (5, 8)),
]
for g,r,yrange in GRs:
for fitter in fitters:
print 'Doing {}@{}...'.format(g,r)
series = data.get_one_series(g,r)
theta,_,_,_ = fitter.fit(series.ages, series.single_expression)
fig = plot_one_series(series, fitter.shape, theta, yrange)
save_figure(fig,'RP/fit-examples-{}-{}-{}.png'.format(fitter.shape.cache_name(), g,r), under_results=True)
#####################################################
# Comparison for whole pathway
#####################################################
pathway = '17full'
data = data.restrict_pathway(pathway)
fits = [get_all_fits(data,fitter,allow_new_computation=False) for fitter in fitters]
fig = plot_comparison_bar(data, shapes, fits)
save_figure(fig,'RP/sigslope-comparison-bar-{}.png'.format(data.pathway), under_results=True)
fig = plot_comparison_over_R2_score(data, shapes, fits)
save_figure(fig,'RP/sigslope-comparison-vs-R2-{}.png'.format(data.pathway), under_results=True)
fig = plot_comparison_scatter(data,shapes[0],fits[0],shapes[1],fits[1])
save_figure(fig,'RP/scatter-{}-{}-{}.png'.format(shapes[0],shapes[1],pathway), under_results=True)
plt.close('all')
save_figure(fig, filename, b_close=True, under_results=True)
dct_tuples.update(dct_region_tuples)
print_best_improvements(dct_tuples)
tuples = dct_tuples.values()
pairs = [(x[0],x[1]) for x in tuples]
fig = plot_comparison_scatter(pairs,pathway)
filename = join('RP','correlation-diff-scatter-{}.png'.format(pathway))
save_figure(fig, filename, b_close=True, under_results=True)
fig = plot_comparison_bar(tuples)
filename = join('RP','correlation-diff-bar-{}.png'.format(pathway))
save_figure(fig, filename, b_close=True, under_results=True)
fig = plot_comparison_bar(tuples, several_levels=True)
filename = join('RP','correlation-diff-bar-several-levels-{}.png'.format(pathway))
save_figure(fig, filename, b_close=True, under_results=True)
disable_all_warnings()
cfg.verbosity = 1
age_scaler = LogScaler()
shape = Sigslope('sigslope80')
fitter = Fitter(shape, sigma_prior='normal')
pathways = ['cannabinoids', 'serotonin']
for pathway in pathways:
data = GeneData.load('both').restrict_pathway(pathway).scale_ages(age_scaler)
fits = get_all_fits(data, fitter, n_correlation_iterations=4, allow_new_computation=False)
analyze_pathway(pathway, data, fitter, fits)
def get_fits():
data = GeneData.load('both').restrict_pathway('17pathways').scale_ages(age_scaler)
shape = Sigmoid(priors='sigmoid_wide')
fitter = Fitter(shape, sigma_prior='normal')
fits = get_all_fits(data, fitter)
return fits
tuples = dct_tuples.values()
pairs = [(x[0], x[1]) for x in tuples]
fig = plot_comparison_scatter(pairs, pathway)
filename = join('RP', 'correlation-diff-scatter-{}.png'.format(pathway))
save_figure(fig, filename, b_close=True, under_results=True)
fig = plot_comparison_bar(tuples)
filename = join('RP', 'correlation-diff-bar-{}.png'.format(pathway))
save_figure(fig, filename, b_close=True, under_results=True)
fig = plot_comparison_bar(tuples, several_levels=True)
filename = join(
'RP', 'correlation-diff-bar-several-levels-{}.png'.format(pathway))
save_figure(fig, filename, b_close=True, under_results=True)
disable_all_warnings()
cfg.verbosity = 1
age_scaler = LogScaler()
shape = Sigslope('sigslope80')
fitter = Fitter(shape, sigma_prior='normal')
pathways = ['cannabinoids', 'serotonin']
for pathway in pathways:
data = GeneData.load('both').restrict_pathway(pathway).scale_ages(
age_scaler)
fits = get_all_fits(data,
fitter,
n_correlation_iterations=4,
allow_new_computation=False)
analyze_pathway(pathway, data, fitter, fits)
import setup
import config as cfg
from load_data import GeneData
from shapes.sigmoid import Sigmoid
from fitter import Fitter
from all_fits import get_all_fits, iterate_fits
from scalers import LogScaler
cfg.verbosity = 1
age_scaler = LogScaler()
pathway = 'serotonin'
data = GeneData.load('both').restrict_pathway(pathway).scale_ages(age_scaler)
fitter = Fitter(Sigmoid(priors=None))
fits = get_all_fits(data, fitter)
extreme = [(g, r) for dsname, g, r, fit in iterate_fits(
fits, R2_threshold=0.5, return_keys=True) if abs(fit.theta[0]) > 100]
def save_fits_and_create_html(data,
fitter,
fits=None,
basedir=None,
do_genes=True,
do_series=True,
do_hist=True,
do_html=True,
only_main_html=False,
k_of_n=None,
use_correlations=False,
correlations=None,
show_change_distributions=False,
exons_layout=False,
html_kw=None,
figure_kw=None):
if fits is None:
fits = get_all_fits(data, fitter, k_of_n)
if basedir is None:
basedir = join(results_dir(), fit_results_relative_path(data, fitter))
if use_correlations:
basedir = join(basedir, 'with-correlations')
if html_kw is None:
html_kw = {}
if figure_kw is None:
figure_kw = {}
print 'Writing HTML under {}'.format(basedir)
ensure_dir(basedir)
gene_dir = 'gene-subplot'
exons_dir = 'exons_subplot_series' if cfg.exons_plots_from_series else 'exons_subplot'
series_dir = 'gene-region-fits'
correlations_dir = 'gene-correlations'
scores_dir = 'score_distributions'
if do_genes and not only_main_html: # relies on the sharding of the fits respecting gene boundaries
plot_and_save_all_genes(data, fitter, fits, join(basedir, gene_dir),
show_change_distributions)
if do_series and not only_main_html:
plot_and_save_all_series(data, fitter, fits, join(basedir, series_dir),
use_correlations, show_change_distributions,
exons_layout, figure_kw)
if exons_layout and not only_main_html:
if cfg.exons_plots_from_series:
plot_and_save_all_exons_from_series(fits, join(basedir, exons_dir),
join(basedir, series_dir))
else:
plot_and_save_all_exons(data, fitter, fits,
join(basedir, exons_dir))
if do_hist and k_of_n is None and not only_main_html:
create_score_distribution_html(fits, use_correlations,
join(basedir, scores_dir))
if do_html and k_of_n is None:
link_to_correlation_plots = use_correlations and correlations is not None
if link_to_correlation_plots and not only_main_html:
plot_and_save_all_gene_correlations(
data, correlations, join(basedir, correlations_dir))
dct_pathways = load_17_pathways_breakdown()
pathway_genes = set.union(*dct_pathways.values())
data_genes = set(data.gene_names)
missing = pathway_genes - data_genes
b_pathways = len(missing) < len(
pathway_genes
) / 2 # simple heuristic to create pathways only if we have most of the genes (currently 61 genes are missing)
create_html(data,
fitter,
fits,
basedir,
gene_dir,
exons_dir,
series_dir,
scores_dir,
correlations_dir=correlations_dir,
use_correlations=use_correlations,
link_to_correlation_plots=link_to_correlation_plots,
b_pathways=b_pathways,
exons_layout=exons_layout,
**html_kw)
import setup
import config as cfg
from load_data import GeneData
from shapes.sigmoid import Sigmoid
from fitter import Fitter
from all_fits import get_all_fits, iterate_fits
from scalers import LogScaler
cfg.verbosity = 1
age_scaler = LogScaler()
pathway = 'serotonin'
data = GeneData.load('both').restrict_pathway(pathway).scale_ages(age_scaler)
fitter = Fitter(Sigmoid(priors=None))
fits = get_all_fits(data,fitter)
extreme = [(g,r) for dsname,g,r,fit in iterate_fits(fits, R2_threshold=0.5, return_keys=True) if abs(fit.theta[0]) > 100]
ax.bar(bin_edges[:-1], probs, width=width, color='b')
ax.set_xlabel('z score', fontsize=fontsize)
ax.set_ylabel('probability', fontsize=fontsize)
ax.tick_params(axis='both', labelsize=fontsize)
return fig
cfg.verbosity = 1
age_scaler = LogScaler()
pathway = '17full'
data = GeneData.load('both').restrict_pathway(pathway).scale_ages(age_scaler)
data_shuffled = GeneData.load('both').restrict_pathway(pathway).scale_ages(age_scaler).shuffle()
shape = Sigmoid('sigmoid_wide')
fitter = Fitter(shape,sigma_prior='normal')
fits = get_all_fits(data,fitter,allow_new_computation=False)
fits_shuffled = get_all_fits(data_shuffled,fitter,allow_new_computation=False)
R2_pairs = [(fit.LOO_score,fit2.LOO_score) for fit,fit2 in iterate_fits(fits,fits_shuffled)]
R2 = np.array([r for r,r_shuffled in R2_pairs])
R2_shuffled = np.array([r_shuffled for r,r_shuffled in R2_pairs])
name = '{}-{}'.format(data.pathway,shape.cache_name())
fig = plot_score_distribution(R2,R2_shuffled)
save_figure(fig,'RP/R2-distribution-{}.png'.format(name), under_results=True, b_close=True)
mu_shuffled = np.mean(R2_shuffled)
std_shuffled = np.std(R2_shuffled)
z_scores = (R2-mu_shuffled)/std_shuffled
fig = plot_z_scores(z_scores)
save_figure(fig,'RP/R2-z-scores-{}.png'.format(name), under_results=True, b_close=True)
