if True:
for i, gp in enumerate(groups):
yzer.piechart([sum(gp['naive_foxo1_tag_count'] >= min_thresh),
sum(gp['naive_foxo1_tag_count'] < min_thresh)],
['With Foxo1', 'Without Foxo1'],
title='Co-occurrence with Foxo1- ' + labels[i],
save_dir=save_path, show_plot=False)
yzer.piechart([sum(gp['lcmv_d12_foxo1_tag_count'] >= min_thresh),
sum(gp['lcmv_d12_foxo1_tag_count'] < min_thresh)],
['With Foxo1', 'Without Foxo1'],
title='Co-occurrence with LCMV d12 Foxo1- ' +
labels[i],
save_dir=save_path, show_plot=False)
yzer.histogram(gp['naive_foxo1_tag_count'].tolist(), bins=20,
title='Foxo1 peak tag count distribution- ' +
labels[i],
xlabel='Tag count in Foxo1 peak',
ylabel='Number of peaks',
save_dir=save_path, show_plot=False)
if True:
yzer.boxplot([gp[gp['naive_foxo1_tag_count'] > 0]['naive_foxo1_tag_count']
for gp in groups],
labels,
title='ATAC-seq regions with Foxo1 peaks by group',
ylabel='Foxo1 peak tag count', save_dir=save_path,
show_plot=False)
# TCF1
if True:
for i, gp in enumerate(groups):
yzer.piechart([sum(gp['tcf1_tag_count'] >= min_thresh),
shuffled = data['domain_id'].values.copy()
shuffle(shuffled)
data['shuffled_domain_id'] = shuffled
grouped = data.groupby(by='domain_id', as_index=False).mean()
shuffled_grouped = data.groupby(by='shuffled_domain_id',
as_index=False).mean()
grouped = grouped[grouped['domain_id'] != 0]
shuffled_grouped = shuffled_grouped[
shuffled_grouped['shuffled_domain_id'] != 0]
ax = yzer.histogram(grouped[dex_kla_key],
bins=50,
label='Replicate {0} Data'.format(rep),
show_legend=False,
show_plot=False)
ax = yzer.histogram(
shuffled_grouped[dex_kla_key],
bins=50,
title='Dex+KLA over KLA LFC {0} by HiC Domain'.format(rep),
xlabel='Mean Dex+KLA log fold change for transcripts up in KLA',
ylabel='Count of Domains',
color='black',
fill=False,
label='Shuffled Data'.format(rep),
show_legend=True,
save_dir=img_dirpath,
show_plot=False,
grouped = grouped[grouped['up_in_kla'] >= 1]
shuffled_grouped = shuffled_grouped[shuffled_grouped['up_in_kla'] >= 1]
grouped['kla_ratio'] = grouped['up_in_kla'] / grouped['count']
grouped = grouped.sort(['kla_ratio']).reset_index(drop=True)
grouped['idx'] = grouped.index
shuffled_grouped['kla_ratio'] = shuffled_grouped[
'up_in_kla'] / shuffled_grouped['count']
shuffled_grouped = shuffled_grouped.sort(['kla_ratio'
]).reset_index(drop=True)
shuffled_grouped['idx'] = shuffled_grouped.index
ax = yzer.histogram(grouped['kla_ratio'],
bins=50,
label='Replicate {0} Data'.format(rep),
show_legend=False,
show_plot=False)
ax = yzer.histogram(
shuffled_grouped['kla_ratio'],
bins=50,
title='Percent of transcripts up in KLA {0} by HiC Domain'.format(
rep),
ylabel='Count of domains',
xlabel='Percent of transcripts up in KLA',
color='black',
fill=False,
label='Shuffled Data'.format(rep),
show_legend=True,
save_dir=img_dirpath,
print 'Percentage: ', len(subset) / len(refseq)
print 'mRNA: ', sum(subset['type'] == 'mRNA'), sum(
subset['type'] == 'mRNA') / len(subset)
print 'rRNA: ', sum(subset['type'] != 'mRNA'), sum(
subset['type'] != 'mRNA') / len(subset)
# Note that some transcripts encompass two very close refseq genes; we're filtering those out for now.
mrna_with_runoff = refseq_with_runoff[
(refseq_with_runoff['type'] == 'mRNA')
& (refseq_with_runoff['refseq'] == 't')
& (refseq_with_runoff['percent_covered'] < 1.5)]
mrna_no_runoff = refseq_no_runoff[
(refseq_no_runoff['type'] == 'mRNA')
& (refseq_no_runoff['refseq'] == 't')
& (refseq_no_runoff['percent_covered'] < 1.5)]
if False:
yzer.histogram(mrna_with_runoff['percent_covered'])
yzer.histogram(mrna_no_runoff['percent_covered'])
relationships = ['is contained by', 'contains', 'overlaps with']
with_runoff_counts = [
sum(mrna_with_runoff['relationship'] == rel)
for rel in relationships
]
no_runoff_counts = [
sum(mrna_no_runoff['relationship'] == rel)
for rel in relationships
]
yzer.piechart(with_runoff_counts, labels=relationships)
yzer.piechart(no_runoff_counts, labels=relationships)
if True:
dirpath, 'Figures', 'me2_atac_overlaps')
yzer.piechart([len(atac_only), len(atac_me2)],
['ATAC only', 'ATAC with H3K4me2'],
title='ATAC-seq region overlaps',
save_dir=save_path)
yzer.piechart([len(me2_only), len(me2_atac)],
['H3K4me2 only', 'H3K4me2 with ATAC'],
title='H3K4me2 overlaps',
save_dir=save_path)
yzer.boxplot([atac_only['tag_count'], atac_me2['tag_count']],
['ATAC only', 'ATAC with H3K4me2'],
title='ATAC-seq tag counts by H3K4me2 overlap',
xlabel='Group', ylabel='Peak tag count',
save_dir=save_path)
yzer.boxplot([me2_only['tag_count'], me2_atac['tag_count']],
['H3K4me2 only', 'H3K4me2 with ATAC'],
title='H3K4me2 tag counts by ATAC-seq overlap',
xlabel='Group', ylabel='Peak tag count',
save_dir=save_path)
yzer.histogram(atac_only['tag_count'].tolist(), bins=20,
title='ATAC-seq-only peak tag count distribution',
xlabel='Tag count in peak', ylabel='Number of peaks',
save_dir=save_path)
yzer.histogram(me2_only['tag_count'].tolist(), bins=20,
title='H3K4me2-only peak tag count distribution',
xlabel='Tag count in peak', ylabel='Number of peaks',
save_dir=save_path)
