indir = os.path.join(DATA_DIR, 'rnaseq', 'wtchg_p160704')
lanedirs = [
os.path.join(indir, '161219_K00198_0151_BHGYHTBBXX'),
os.path.join(indir, '161222_K00198_0152_AHGYG3BBXX'),
]
metafiles = [os.path.join(d, 'sources.csv') for d in lanedirs]
countdirs = [os.path.join(d, 'star_alignment_mouse') for d in lanedirs]
obj2 = rnaseq_data.all_samples_multilane_loader(countdirs,
metafiles,
source='star',
annotate_by='Ensembl Gene ID',
samples=('ICb1078', 'ICb1487'))
obj = rnaseq_data.MultipleBatchLoader([obj1, obj2])
data = obj.data.loc[obj.data.index.str.contains('ENS')]
cpm = data.divide(obj.meta.read_count, axis=1) * 1e6
cpm = cpm.astype(float)
# only keep abundant genes: results in around 12,000 remaining genes
keep = (cpm > 1).sum(axis=1) > 5
fig = plt.figure(figsize=(10, 10))
ax = fig.add_subplot(111)
ax = corr.plot_correlation_coefficient_array(data.loc[keep], vmin=0.6, ax=ax)
plt.setp(ax.xaxis.get_ticklabels(), rotation=90)
fig.tight_layout()
# 2) What are the correlation coeffs when we look at individual lanes?
# Only include the endogenous NSC in the endogenous medium
])
obj52564 = rnaseq_data.gse52564(annotate_by='Ensembl Gene ID',
samples=[
'Astrocyte1',
'Astrocyte2',
'Neuron1',
'Neuron2',
'OPC1',
'OPC2',
])
obj43916 = rnaseq_data.gse43916(annotate_by='Ensembl Gene ID',
samples=['NSCs'])
obj86248 = rnaseq_data.gse86248(annotate_by='Ensembl Gene ID')
obj36114 = rnaseq_data.gse36114(annotate_by='Ensembl Gene ID')
obj_all = rnaseq_data.MultipleBatchLoader(
[obj, obj64411, obj52564, obj43916, obj86248, obj36114])
data_all = obj_all.data.loc[obj_all.data.index.str.contains('ENS')]
meta_all = obj_all.meta
cpm_all = data_all.divide(data_all.sum(axis=0), axis=1) * 1e6
keep = (cpm_all > .5).sum(axis=1) > 5
the_dat = np.log2(data_all.loc[keep] + 1)
fig = plt.figure(figsize=(10, 10))
ax = fig.add_subplot(111)
ax = corr.plot_correlation_coefficient_array(the_dat, vmin=0.4, ax=ax)
plt.setp(ax.xaxis.get_ticklabels(), rotation=90)
fig.tight_layout()
fig.savefig(os.path.join(outdir, 'corr_coeff.png'), dpi=200)
subgroups = {
'RTK I': ['019', '030', '031'],
'RTK II': ['017', '050', '054'],
}
intersecter = lambda x, y: set(x).intersection(y)
unioner = lambda x, y: set(x).union(y)
# Load RNA-Seq from STAR
rnaseq_obj = rnaseq_data.load_by_patient(pids, annotate_by='Ensembl Gene ID')
# load additional references if required
h9_obj = rnaseq_data.gse61794(annotate_by='Ensembl Gene ID')
h1_obj = rnaseq_data.gse38993(annotate_by='Ensembl Gene ID')
rnaseq_obj = rnaseq_data.MultipleBatchLoader([rnaseq_obj, h1_obj, h9_obj])
# discard unmapped, etc
rnaseq_obj.data = rnaseq_obj.data.loc[rnaseq_obj.data.index.str.contains('ENSG')]
rnaseq_obj.meta = rnaseq_obj.meta.loc[~rnaseq_obj.meta.index.str.contains('PSC')]
rnaseq_obj.meta = rnaseq_obj.meta.loc[~rnaseq_obj.meta.index.str.contains('fibroblast')]
rnaseq_obj.data = rnaseq_obj.data.loc[:, rnaseq_obj.meta.index]
# load RNA-Seq from Salmon (for normalised comparison)
# disabled for now
if False:
salmon_dat = rnaseq_data.load_salmon_by_patient_id(pids)
idx = salmon_dat.index.str.replace(r'.[0-9]+$', '')
salmon_dat.index = idx
fn = os.path.join(LOCAL_DATA_DIR, 'reference_genomes', 'human', 'ensembl', 'GRCh38.p10.release90',
'gene_to_transcript.txt')
if njob != 1:
pool = mp.Pool(njob)
# Load RNA-Seq from STAR
rnaseq_obj = rnaseq_data.load_by_patient(pids,
annotate_by='Ensembl Gene ID')
# load additional references if required
refs = [('H1', rnaseq_data.gse38993(annotate_by='Ensembl Gene ID')),
('H9',
rnaseq_data.gse61794(annotate_by='Ensembl Gene ID',
collapse_replicates=False))]
all_refs = [t[0] for t in refs] + ['GIBCO']
rnaseq_obj = rnaseq_data.MultipleBatchLoader([rnaseq_obj] +
[t[1] for t in refs])
# only keep gene counts
rnaseq_obj.data = rnaseq_obj.data.loc[rnaseq_obj.data.index.str.contains(
'ENSG')]
# discard iPSC
rnaseq_obj.meta = rnaseq_obj.meta.loc[~rnaseq_obj.meta.index.str.
contains('PSC')]
rnaseq_obj.meta = rnaseq_obj.meta.loc[~rnaseq_obj.meta.index.str.
contains('NHF1-hTERT')]
rnaseq_obj.meta = rnaseq_obj.meta.loc[~rnaseq_obj.meta.index.str.
contains('fibroblast')]
rnaseq_obj.meta = rnaseq_obj.meta.loc[~rnaseq_obj.meta.index.str.
contains('ESC')]