def load_data_no_impute_event(etype, event_id, map_event_to_file): # , sf_interest):
df_list = list()
path = map_event_to_file[etype]
data = h5py.File(path, 'r')
gene_idx = data['gene_idx'][event_id].astype('int')
ensg_idx = data['gene_names'][gene_idx]
lookup = names.get_lookup_complete()
ensg_name = names.get_ID(ensg_idx, lookup=lookup)
psi = data['psi'][:, event_id]
columns = [_encode_event_name(re.sub(r'-', '_', ensg_name), etype, event_id)]
strains = sf_utils.clean_strain(data['strains'][:])
df_list.append(pd.DataFrame(psi, index=strains, columns=columns))
df = pd.concat(df_list, axis=1)
return df
def get_sf_interest(v=False):
gene_name_list = ['PKM', 'MAX', 'NUMB', 'MKNK2', 'BIN1',
'RPS6KB1', 'BCL2L1', 'APC', 'PTEN',
'KLF6', 'CD44', 'CCK2', 'GHRH',
'CDKN2A', 'KIT']
sf_interest = dict()
lookup = names.get_lookup_complete()
for gene_name in gene_name_list:
gene_ensg = names.get_ID(gene_name, which='ensembl', lookup=lookup)
if not gene_ensg.upper().startswith('ENSG'):
if v: print "WARNING: %s -> %s is unexpected and ignored" %(gene_name, gene_ensg)
continue
sf_interest[gene_name] = gene_ensg
if v:
print "Genes of interest:"
display(sf_interest)
return sf_interest
merged_ids = merged_ids[s_idx, :]
genes = genes[s_idx]
print("Writing output files")
merged_bin = (merged > 4.5).astype('int').astype('str')
merged_bin = sp.r_[strains[sp.newaxis, :], merged_bin]
merged_bin = sp.c_[sp.r_[['gene_id'], genes], merged_bin]
merged_median = sp.median(merged, axis=1).astype('str')
merged_median = sp.append('median_dev', merged_median)
merged_median = sp.c_[
sp.append('gene_id', sp.array([x.split('.')[0] for x in genes])),
merged_median]
merged_median = sp.c_[sp.append(
'gene_name',
sp.array([names.get_ID(x.split('.')[0], lookup) for x in genes])),
merged_median]
merged_str = merged.astype('str')
merged_str = sp.r_[strains[sp.newaxis, :], merged_str]
merged_str = sp.c_[sp.r_[['gene_id'], genes], merged_str]
merged_ids_str = merged_ids.astype('str')
merged_ids_str = sp.r_[strains[sp.newaxis, :], merged_ids_str]
merged_ids_str = sp.c_[sp.r_[['gene_id'], genes], merged_ids_str]
if not os.path.exists(os.path.join(basedir, outfile_pat[p])):
sp.savetxt(os.path.join(basedir, outfile_pat[p]),
merged_str,
fmt='%s',
delimiter='\t')
alt_projects = sp.array(alt_projects) var_pos = sp.array(var_pos) alt_files = sp.array(alt_files) ref_files = sp.array(ref_files) k_idx = sp.where(sp.in1d(keep_idx, conf_idx))[0] keep_idx = keep_idx[k_idx] delta_psi = delta_psi[k_idx] alt_donors = alt_donors[k_idx] ref_donors = ref_donors[k_idx] alt_projects = alt_projects[k_idx] var_pos = var_pos[k_idx] alt_files = alt_files[k_idx] ref_files = ref_files[k_idx] event_pos = sp.array([EV['event_pos'][i, :] for i in keep_idx]) gene_idx = sp.array([EV['gene_idx'][i] for i in keep_idx], dtype='int') gene_strand = sp.array([EV['gene_strand'][i] for i in gene_idx]) gene_ids = sp.array([EV['gene_names'][i] for i in gene_idx]) gene_names = sp.array([un.get_ID(_, lookup=lookup) for _ in gene_ids]) is_coding = sp.array(['coding' if _ in coding_genes else 'non-coding' for _ in gene_ids]) event_chr = sp.array([EV['gene_chr'][i] for i in gene_idx]) s_idx = sp.argsort(delta_psi)[::-1] # 0 1 2 3--8 9 10 11 12 13 14 15 16 17 18 res = sp.c_[var_pos, keep_idx, event_chr, event_pos, gene_ids, gene_names, delta_psi, alt_donors, ref_donors, alt_projects, alt_files, ref_files, gene_strand, is_coding][s_idx, :] sp.savetxt(os.path.join(BASEDIR_AS, 'alternative_splicing', 'exonization_candidates_C2.txt'), res, fmt='%s', delimiter='\t') EV.close()
strains = sp.array([x.split('.')[1] for x in IN['strains'][:]])
### apply whitelist
if use_wl:
whitelist = sp.loadtxt(paths.whitelist, delimiter='\t', dtype='str')
whitelist = sp.array([x.split('.')[1] for x in whitelist])
widx = sp.in1d(strains, whitelist)
strains = strains[widx]
else:
widx = sp.arange(strains)
ctypes = sp.array([ct_dict[x] for x in strains])
is_tumor = sp.array([is_tumor_dict[x] for x in strains])
ctypes_u, ctypes_cnt = sp.unique(ctypes, return_counts=True)
gene_ids = IN['gene_names'][:]
gene_names = sp.array([nm.get_ID(x, lookup=lookup) for x in gene_ids])
gene_idx = IN['gene_idx'][:].astype('int')
### only keep cancer types where we have at least 100 samples
k_idx = sp.where(ctypes_cnt >= 100)[0]
ctypes_u = ctypes_u[k_idx]
outliers[et] = dict([(x, []) for x in ctypes_u])
chunks = IN['psi'].chunks
for cc, c in enumerate(range(0, IN['psi'].shape[1], chunks[1])):
sys.stdout.write('%i/%i - outliers: %i\n' % (cc, IN['psi'].shape[1] / chunks[1], outlier_cnt))
sys.stdout.flush()
cidx = sp.arange(c, min(c + chunks[1], IN['psi'].shape[1]))
tmp = IN['psi'][:, cidx][widx, :]
tmp_gt = IN_GT['psi'][:, cidx]
rows = alt_donors.shape[0] + 1
height_ratios = [3 for _ in range(rows)]
height_ratios[-1] = 1
gs = gridspec.GridSpec(rows, 1, height_ratios=height_ratios)
fig = plt.figure(figsize=(16, 3 * rows), dpi=200)
#fig.suptitle('%s in %s (Variant: %s:%s)' % (results[i, 6], results[i, 7],results[i, 0], results[i, 2]), fontsize=12, y=1.03)
for d, donor in enumerate(alt_donors):
### plot coverage profile
ax = fig.add_subplot(gs[d, 0])
cax = []
labels = []
#ax.set_title('%s - Gene: %s (%s) - SNV: %s - in frame: %s' % (results[i, 14], un.get_ID(results[i, 9], lookup=lookup), results[i, 17], results[i, 0], in_frame), fontsize=10)
ax.set_title('Gene: %s -in frame: %s' %
(un.get_ID(results[i, 8], lookup=lookup), in_frame),
fontsize=10)
cax.append(
cv.cov_from_bam(event_chr,
event_pos.min(),
event_pos.max(), [files_alt[d]],
color_cov='r',
ax=ax,
intron_cnt=True,
color_intron_edge='r',
log=('--log' in sys.argv),
intron_filter=intron_filter,
return_legend_handle=True,
label='Alt: %s' % donor))
labels.append('Alt: %s' % donor)
if files_ref[d] != 'NA':
event_chr = str(results[i, 2])
rows = alt_donors.shape[0] + 1
height_ratios = [3 for _ in range(rows)]
height_ratios[-1] = 1
gs = gridspec.GridSpec(rows, 1, height_ratios=height_ratios)
fig = plt.figure(figsize = (8, 3 * rows), dpi=200)
#fig.suptitle('%s in %s (Variant: %s:%s)' % (results[i, 6], results[i, 7],results[i, 0], results[i, 2]), fontsize=12, y=1.03)
for d, donor in enumerate(alt_donors):
### plot coverage profile
ax = fig.add_subplot(gs[d, 0])
cax = []
labels = []
ax.set_title('%s - Gene: %s (%s) - SNV: %s - in frame: %s' % (results[i, 14], un.get_ID(results[i, 9], lookup=lookup), results[i, 17], results[i, 0], in_frame), fontsize=10)
cax.append(cv.cov_from_bam(event_chr, event_pos.min(), event_pos.max(), [files_alt[d]], color_cov='r', ax=ax, intron_cnt=True, color_intron_edge='r', log=('--log' in sys.argv), intron_filter=intron_filter, return_legend_handle=True, label='Alt: %s' % donor))
labels.append('Alt: %s' % donor)
if files_ref[d] != 'NA':
cax.append(cv.cov_from_bam(event_chr, event_pos.min(), event_pos.max(), [files_ref[d]], color_cov='0.0', ax=ax, intron_cnt=True, color_intron_edge='0.0', log=('--log' in sys.argv), intron_filter=intron_filter, return_legend_handle=True, label='Ref: %s' % ref_donors[d]))
labels.append('Ref: %s' % ref_donors[d])
plt.legend(cax, labels, fontsize=10)
for vp in var_pos:
ax.plot(vp, 0, 'bo', markersize=1)
ylim = ax.get_ylim()
xlim = ax.get_xlim()
xspan = xlim[1] - xlim[0]
yspan = ylim[1] - ylim[0]
for vp in var_pos:
ax.arrow(vp, yspan * 0.2, 0, yspan * -0.15, head_width=0.01*xspan, head_length=0.01*yspan, fc='b')
x_range = ax.get_xlim()
rows = alt_donors.shape[0] + 1
height_ratios = [3 for _ in range(rows)]
height_ratios[-1] = 1
gs = gridspec.GridSpec(rows, 1, height_ratios=height_ratios)
fig = plt.figure(figsize=(16, 3 * rows), dpi=200)
#fig.suptitle('%s in %s (Variant: %s:%s)' % (results[i, 6], results[i, 7],results[i, 0], results[i, 2]), fontsize=12, y=1.03)
for d, donor in enumerate(alt_donors):
### plot coverage profile
ax = fig.add_subplot(gs[d, 0])
cax = []
labels = []
ax.set_title('%s - Gene: %s (%s) - SNV: %s - in frame: %s' %
(results[i, 14], un.get_ID(results[i, 9], lookup=lookup),
results[i, 17], results[i, 0], in_frame),
fontsize=10)
cax.append(
cv.cov_from_bam(event_chr,
event_pos.min(),
event_pos.max(), [files_alt[d]],
color_cov='r',
ax=ax,
intron_cnt=True,
color_intron_edge='r',
log=('--log' in sys.argv),
intron_filter=intron_filter,
return_legend_handle=True,
label='Alt: %s' % donor))
labels.append('Alt: %s' % donor)