def test_feature_matrix(self):
# No segments
features = fstream([(5,15,'gene1'),(30,40,'gene2')], fields=['start','end','name'])
scores1 = fstream([(10,15,6.),(30,40,6.)], fields=['start','end','score'])
scores2 = fstream([(5,15,2.)], fields=['start','end','score'])
feat, res = feature_matrix([scores1,scores2],features)
self.assertListEqual(list(feat),['gene1','gene2'])
assert_almost_equal(res, numpy.array([[3,2],[6,0]]))
# Segmenting each feature in 3 parts
features = fstream([(5,15,'gene1'),(30,40,'gene2')], fields=['start','end','name'])
scores1 = fstream([(10,15,6.),(30,40,6.)], fields=['start','end','score'])
scores2 = fstream([(5,15,2.)], fields=['start','end','score'])
feat, res = feature_matrix([scores1,scores2],features, segment=True, nbins=3)
assert_almost_equal(res, numpy.array([[[0,2],[2,2],[6,2]],
[[6,0],[6,0],[6,0]]]))
def test_feature_matrix(self):
# No segments
features = fstream([(5, 15, 'gene1'), (30, 40, 'gene2')],
fields=['start', 'end', 'name'])
scores1 = fstream([(10, 15, 6.), (30, 40, 6.)],
fields=['start', 'end', 'score'])
scores2 = fstream([(5, 15, 2.)], fields=['start', 'end', 'score'])
feat, res = feature_matrix([scores1, scores2], features)
self.assertListEqual(list(feat), ['gene1', 'gene2'])
assert_almost_equal(res, numpy.array([[3, 2], [6, 0]]))
# Segmenting each feature in 3 parts
features = fstream([(5, 15, 'gene1'), (30, 40, 'gene2')],
fields=['start', 'end', 'name'])
scores1 = fstream([(10, 15, 6.), (30, 40, 6.)],
fields=['start', 'end', 'score'])
scores2 = fstream([(5, 15, 2.)], fields=['start', 'end', 'score'])
feat, res = feature_matrix([scores1, scores2],
features,
segment=True,
nbins=3)
assert_almost_equal(
res,
numpy.array([[[0, 2], [2, 2], [6, 2]], [[6, 0], [6, 0], [6, 0]]]))
def __call__(self, **kw):
def make_X_labels(X,start,end,strand,down,up):
flen = end-start
i0 = (list(where(X == 0)[0])+[-1])[0]+1
i1 = (list(where(X == 1)[0])+[len(X)-1])[0]+1
i2 = len(X)-i1
istep = 0.5/(i1-i0)
if down < 1: down *= flen
if up < 1: up *= flen
Xup = (array(range(-i0,0))+.5)*up/i0
Xb = (X[i0:i1]+istep)*flen
Xdown = flen+(array(range(i2))+.5)*down/i2
if strand is None or strand > 0: return start+concatenate([Xup,Xb,Xdown])
else: return end-concatenate([Xup,Xb,Xdown])
chrmeta = "guess"
features = track(kw.get('features'), chrmeta=chrmeta)
#signals = kw.get('SigMulti',{}).get('signals', [])
signals = kw.get('signals', [])
if not isinstance(signals, list): signals = [signals]
signals = [track(sig) for sig in signals]
snames = [sig.name for sig in signals]
labels = None
data = None
upstr = _upstr
downstr = _downstr
if kw.get("upstream") is not None:
_up = int(kw["upstream"])
if _up > 50: upstr = (_up,5)
elif _up > 0: upstr = (_up,1)
else: upstr = (0,0)
if kw.get("downstream") is not None:
_down = int(kw["downstream"])
if _down > 50: downstr = (_down,5)
elif _down > 0: downstr = (_down,1)
else: downstr = (0,0)
if kw.get("nbins") is not None: nbins = max(1,int(kw["nbins"]))
else: nbins = _nbins
if kw.get("noclust") is not None:
noclust = str(kw["noclust"]).lower() in ['1','true','t','on']
else:
noclust = False
try:
ymin = float(kw.get('ymin'))
except (ValueError, TypeError):
ymin = None
try:
ymax = float(kw.get('ymax'))
except (ValueError, TypeError):
ymax = None
for chrom in features.chrmeta:
if 'name' in features.fields: _fread = features.read(chrom)
else: _fread = add_name_field(features.read(chrom))
_l, _d = feature_matrix([s.read(chrom) for s in signals], _fread,
segment=True, nbins=nbins,
upstream=upstr, downstream=downstr)
if _d.size == 0:
continue
if data is None:
labels = _l
data = _d
else:
labels = concatenate((labels, _l))
data = vstack((data, _d))
outf = str(kw.get('output'))
if outf not in output_list:
outf = output_list[0]
pdf = self.temporary_path(fname='plot_features.pdf')
if outf == 'archive':
tarname = self.temporary_path(fname='plot_features.tar.gz')
tarfh = tarfile.open(tarname, "w:gz")
if data is None:
raise ValueError("No data")
mode = kw.get('mode', 0)
if str(mode) in [str(x[0]) for x in plot_types]:
mode = int(mode)
X = array(range(-upstr[1]+1,nbins+downstr[1]+1))/(1.0*nbins)
if mode in plot_types[0]: #heatmap
new = True
if 'name' in features.fields: _fread = features.read(fields=['chr','start','end','name'])
else: _fread = add_name_field(features.read(fields=['chr','start','end']))
order = [where(labels == feat[3])[0][0] for feat in _fread]
for n in range(data.shape[-1]-1):
heatmap(data[order, :, n], output=pdf, new=new, last=False,
rows=labels[order], columns=X, main=snames[n],
orderRows=not(noclust), orderCols=False,
ymin=ymin, ymax=ymax)
new = False
heatmap(data[order, :, -1], output=pdf, new=new, last=True,
rows=labels[order], columns=X, main=snames[-1],
orderRows=not(noclust), orderCols=False,
ymin=ymin, ymax=ymax)
if outf == 'archive':
for n,sn in enumerate(snames):
_datf = self.temporary_path(fname=sn+"_data.txt")
with open(_datf,"w") as dff:
dff.write("\t".join([""]+[str(x) for x in X])+"\n")
for k in order:
dff.write("\t".join([labels[k]]+[str(x) for x in data[k, :, n]])+"\n")
tarfh.add(_datf,arcname=os.path.basename(_datf))
elif mode in plot_types[1]: #average lineplot
Y = data.mean(axis=0)
if ymin is None: ymin = min([x.min() for x in Y]+[0])
if ymax is None: ymax = max([x.max() for x in Y])
lineplot(X, [Y[:, n] for n in range(data.shape[-1])],
output=pdf, new=True, last=True, legend=snames,
ylim=(ymin,ymax))
if outf == 'archive':
_datf = self.temporary_path(fname="lineplot_data.txt")
with open(_datf,"w") as dff:
dff.write("\t".join([""]+[str(x) for x in X])+"\n")
for n,sn in enumerate(snames):
dff.write("\t".join([sn]+[str(x) for x in Y[:, n]])+"\n")
tarfh.add(_datf,arcname=os.path.basename(_datf))
elif mode in plot_types[2]: #mosaic
mfrow = [4,3]
nplot = min(data.shape[0], max_pages*mfrow[0]*mfrow[1])
if ymin is None: ymin = min([data.min(),0])
if ymax is None: ymax = data.max()
_f = ['chr','start','end']
_si = None
if 'strand' in features.fields:
_f.append('strand')
_si = 3
if 'name' in features.fields: _fread = features.read(fields=_f+['name'])
else: _fread = add_name_field(features.read(fields=_f))
order = []
for nf,feat in enumerate(_fread):
reg = where(labels == feat[-1])[0][0]
order.append(reg)
X1 = make_X_labels(X, feat[1], feat[2], feat[_si] if _si else None, downstr[0], upstr[0])
xlim = (X1[0],X1[-1])
Y = [data[reg, :, n] for n in range(data.shape[-1])]
if nf == 0:
lineplot(X1, Y, output=pdf, new=True, last=False, mfrow=mfrow,
main=labels[reg], ylim=(ymin,ymax), xlim=xlim)
elif nf < nplot-1:
lineplot(X1, Y, output=pdf, new=False, last=False,
main=labels[reg], ylim=(ymin,ymax), xlim=xlim)
else:
lineplot(X1, Y, output=pdf, new=False, last=True, legend=snames,
main=labels[reg], ylim=(ymin,ymax), xlim=xlim)
break
if outf == 'archive':
for n,sn in enumerate(snames):
_datf = self.temporary_path(fname=sn+"_data.txt")
with open(_datf,"w") as dff:
dff.write("\t".join([""]+[str(x) for x in X])+"\n")
for k in order:
dff.write("\t".join([labels[k]]+[str(x) for x in data[k, :, n]])+"\n")
tarfh.add(_datf,arcname=os.path.basename(_datf))
else:
raise ValueError("Mode not implemented: %s" % mode)
if outf == 'archive':
tarfh.add(pdf,arcname=os.path.basename(pdf))
tarfh.close()
self.new_file(tarname, 'data_archive')
else:
self.new_file(pdf, 'plot_features')
return self.display_time()
def __call__(self, **kw):
def make_X_labels(X, start, end, strand, down, up):
flen = end - start
i0 = (list(where(X == 0)[0]) + [-1])[0] + 1
i1 = (list(where(X == 1)[0]) + [len(X) - 1])[0] + 1
i2 = len(X) - i1
istep = 0.5 / (i1 - i0)
if down < 1: down *= flen
if up < 1: up *= flen
Xup = (array(range(-i0, 0)) + .5) * up / i0
Xb = (X[i0:i1] + istep) * flen
Xdown = flen + (array(range(i2)) + .5) * down / i2
if strand is None or strand > 0:
return start + concatenate([Xup, Xb, Xdown])
else:
return end - concatenate([Xup, Xb, Xdown])
chrmeta = "guess"
features = track(kw.get('features'), chrmeta=chrmeta)
signals = kw.get('SigMulti', {}).get('signals', [])
if not isinstance(signals, list): signals = [signals]
signals = [track(sig) for sig in signals]
snames = [sig.name for sig in signals]
labels = None
data = None
upstr = _upstr
downstr = _downstr
if kw.get("upstream") is not None:
_up = int(kw["upstream"])
if _up > 50: upstr = (_up, 5)
elif _up > 0: upstr = (_up, 1)
else: upstr = (0, 0)
if kw.get("downstream") is not None:
_down = int(kw["downstream"])
if _down > 50: downstr = (_down, 5)
elif _down > 0: downstr = (_down, 1)
else: downstr = (0, 0)
if kw.get("nbins") is not None: nbins = max(1, int(kw["nbins"]))
else: nbins = _nbins
if kw.get("noclust") is not None:
noclust = str(kw["noclust"]).lower() in ['1', 'true', 't', 'on']
else:
noclust = False
try:
ymin = float(kw.get('ymin'))
except (ValueError, TypeError):
ymin = None
try:
ymax = float(kw.get('ymax'))
except (ValueError, TypeError):
ymax = None
for chrom in features.chrmeta:
if 'name' in features.fields: _fread = features.read(chrom)
else: _fread = add_name_field(features.read(chrom))
_l, _d = feature_matrix([s.read(chrom) for s in signals],
_fread,
segment=True,
nbins=nbins,
upstream=upstr,
downstream=downstr)
if _d.size == 0:
continue
if data is None:
labels = _l
data = _d
else:
labels = concatenate((labels, _l))
data = vstack((data, _d))
outf = str(kw.get('output'))
if outf not in output_list:
outf = output_list[0]
pdf = self.temporary_path(fname='plot_features.pdf')
if outf == 'archive':
tarname = self.temporary_path(fname='plot_features.tar.gz')
tarfh = tarfile.open(tarname, "w:gz")
if data is None:
raise ValueError("No data")
mode = kw.get('mode', 0)
if str(mode) in [str(x[0]) for x in plot_types]:
mode = int(mode)
X = array(range(-upstr[1] + 1, nbins + downstr[1] + 1)) / (1.0 * nbins)
if mode in plot_types[0]: #heatmap
new = True
if 'name' in features.fields:
_fread = features.read(fields=['chr', 'start', 'end', 'name'])
else:
_fread = add_name_field(
features.read(fields=['chr', 'start', 'end']))
order = [where(labels == feat[3])[0][0] for feat in _fread]
for n in range(data.shape[-1] - 1):
heatmap(data[order, :, n],
output=pdf,
new=new,
last=False,
rows=labels[order],
columns=X,
main=snames[n],
orderRows=not (noclust),
orderCols=False,
ymin=ymin,
ymax=ymax)
new = False
heatmap(data[order, :, -1],
output=pdf,
new=new,
last=True,
rows=labels[order],
columns=X,
main=snames[-1],
orderRows=not (noclust),
orderCols=False,
ymin=ymin,
ymax=ymax)
if outf == 'archive':
for n, sn in enumerate(snames):
_datf = self.temporary_path(fname=sn + "_data.txt")
with open(_datf, "w") as dff:
dff.write("\t".join([""] + [str(x) for x in X]) + "\n")
for k in order:
dff.write("\t".join(
[labels[k]] + [str(x)
for x in data[k, :, n]]) + "\n")
tarfh.add(_datf, arcname=os.path.basename(_datf))
elif mode in plot_types[1]: #average lineplot
Y = data.mean(axis=0)
if ymin is None: ymin = min([x.min() for x in Y] + [0])
if ymax is None: ymax = max([x.max() for x in Y])
lineplot(X, [Y[:, n] for n in range(data.shape[-1])],
output=pdf,
new=True,
last=True,
legend=snames,
ylim=(ymin, ymax))
if outf == 'archive':
_datf = self.temporary_path(fname="lineplot_data.txt")
with open(_datf, "w") as dff:
dff.write("\t".join([""] + [str(x) for x in X]) + "\n")
for n, sn in enumerate(snames):
dff.write("\t".join([sn] + [str(x)
for x in Y[:, n]]) + "\n")
tarfh.add(_datf, arcname=os.path.basename(_datf))
elif mode in plot_types[2]: #mosaic
mfrow = [4, 3]
nplot = min(data.shape[0], max_pages * mfrow[0] * mfrow[1])
if ymin is None: ymin = min([data.min(), 0])
if ymax is None: ymax = data.max()
_f = ['chr', 'start', 'end']
_si = None
if 'strand' in features.fields:
_f.append('strand')
_si = 3
if 'name' in features.fields:
_fread = features.read(fields=_f + ['name'])
else:
_fread = add_name_field(features.read(fields=_f))
order = []
for nf, feat in enumerate(_fread):
reg = where(labels == feat[-1])[0][0]
order.append(reg)
X1 = make_X_labels(X, feat[1], feat[2],
feat[_si] if _si else None, downstr[0],
upstr[0])
xlim = (X1[0], X1[-1])
Y = [data[reg, :, n] for n in range(data.shape[-1])]
if nf == 0:
lineplot(X1,
Y,
output=pdf,
new=True,
last=False,
mfrow=mfrow,
main=labels[reg],
ylim=(ymin, ymax),
xlim=xlim)
elif nf < nplot - 1:
lineplot(X1,
Y,
output=pdf,
new=False,
last=False,
main=labels[reg],
ylim=(ymin, ymax),
xlim=xlim)
else:
lineplot(X1,
Y,
output=pdf,
new=False,
last=True,
legend=snames,
main=labels[reg],
ylim=(ymin, ymax),
xlim=xlim)
break
if outf == 'archive':
for n, sn in enumerate(snames):
_datf = self.temporary_path(fname=sn + "_data.txt")
with open(_datf, "w") as dff:
dff.write("\t".join([""] + [str(x) for x in X]) + "\n")
for k in order:
dff.write("\t".join(
[labels[k]] + [str(x)
for x in data[k, :, n]]) + "\n")
tarfh.add(_datf, arcname=os.path.basename(_datf))
else:
raise ValueError("Mode not implemented: %s" % mode)
if outf == 'archive':
tarfh.add(pdf, arcname=os.path.basename(pdf))
tarfh.close()
self.new_file(tarname, 'data_archive')
else:
self.new_file(pdf, 'plot_features')
return self.display_time()