def draw_events(canvas,
tree,
coords,
events,
losses,
lossColor=(0, 0, 1),
dupColor=(1, 0, 0),
size=4):
# draw duplications
for node in tree:
x, y = coords[node]
if events[node] == "dup":
canvas.rect(x - size / 2.0,
y - size / 2.0,
size,
size,
fillColor=dupColor,
strokeColor=(0, 0, 0, 0))
# draw losses
losses_per_branch = util.hist_dict([node for node, schild in losses])
for node, nlosses in losses_per_branch.iteritems():
if node.parent == None:
continue
x1 = coords[node.parent][0]
x2, y1 = coords[node]
step = (x2 - x1) / float(nlosses + 1)
for x in util.frange(x1 + step, x2 - (step / 2.0), step):
canvas.line(x, y1 - size, x, y1 + size, color=lossColor)
def make_color_legend(filename, colormap, start, end, step, width=100, height=10, display=False):
from rasmus import util
if filename is None:
filename = util.tempfile(".", "colormap", ".svg")
temp = True
else:
temp = False
s = svg.Svg(util.open_stream(filename, "w"))
s.beginSvg(width, height)
xscale = float(width) / (end + step - start)
for i in util.frange(start, end + step, step):
color = colormap.get(i)
s.rect((i - start) * xscale, 0, step * xscale, height, color, color)
s.endSvg()
s.close()
# display
if display:
os.system("display %s" % filename)
# clean up temp files
if temp:
os.remove(filename)
def make_color_legend(filename,
colormap,
start,
end,
step,
width=100,
height=10,
display=False):
from rasmus import util
if filename is None:
filename = util.tempfile(".", "colormap", ".svg")
temp = True
else:
temp = False
s = svg.Svg(util.open_stream(filename, "w"))
s.beginSvg(width, height)
xscale = float(width) / (end + step - start)
for i in util.frange(start, end + step, step):
color = colormap.get(i)
s.rect((i - start) * xscale, 0, step * xscale, height, color, color)
s.endSvg()
s.close()
# display
if display:
os.system("display %s" % filename)
# clean up temp files
if temp:
os.remove(filename)
def test_plot_prob_bounded_coal(self):
n = 1000
k = 4
t = 800
alltimes = []
# sample times
for i in xrange(5000):
while True:
times = [0]
for j in xrange(k, 1, -1):
times.append(times[-1] + coal.sample_coal(j, n))
if times[-1] >= t:
break
if times[-1] < t:
break
alltimes.append(times)
p = Gnuplot()
for i in range(1, 2):
x, y = distrib([q[i] - q[i-1] for q in alltimes], width=20)
p.plot(x, y, style="lines", xmax=500)
x = list(frange(0, 500, 10))
#for i in range(1, 2): #k):
y2 = [coal.prob_bounded_coal(j, k, n, t) for j in x]
p.plot(x, y2, style="lines", xmax=500)
fequals(y, y2, rel=.05, eabs=.01)
def draw_events(canvas, tree, coords, events, losses,
lossColor=(0, 0, 1),
dupColor=(1, 0, 0),
size=4):
# draw duplications
for node in tree:
x, y = coords[node]
if events[node] == "dup":
canvas.rect(x - size/2.0, y - size/2.0,
size, size, fillColor=dupColor, strokeColor=(0,0,0,0))
# draw losses
losses_per_branch = util.hist_dict([node for node, schild in losses])
for node, nlosses in losses_per_branch.iteritems():
if node.parent == None:
continue
x1 = coords[node.parent][0]
x2, y1 = coords[node]
step = (x2 - x1) / float(nlosses + 1)
for x in util.frange(x1 + step, x2-(step/2.0), step):
canvas.line(x, y1 - size, x, y1 + size, color=lossColor)
def draw_events(self):
# draw duplications
dups = [color(*self.dup_color)]
for node in self.tree:
if self.events[node] == "dup":
dups.append(
zoom_clamp(
shapes.box(node.x - .5, node.y - .5,
node.x + .5, node.y + .5),
link=True, link_type="smaller",
maxx=8, minx=1,
maxy=8, miny=1,
origin=(node.x, node.y),
prezoom=(self.xscale, 1.0)))
# draw losses
losses_per_branch = util.hist_dict([node for node, schild in self.losses])
losses = [color(*self.loss_color)]
for node, nlosses in losses_per_branch.iteritems():
if node.parent == None:
continue
x1 = node.parent.x
x2 = node.x
step = (x2 - x1) / float(nlosses + 1)
for x in util.frange(x1 + step, x2-(step/2.0), step):
losses.append(lines(x, node.y - .2, x, node.y + .2))
return group(group(*dups), group(*losses))
def test_prob_coal_cond_counts(self):
# match against a simpler slower implementation
a = 5
b = 3
t = 2000
n = 1000
for x in frange(0, 2000, 10):
y = coal.prob_coal_cond_counts(x, a, b, t, n)
y2 = coal.prob_coal_cond_counts_simple(x, a, b, t, n)
self.assertAlmostEqual(y, y2)
def test_cdf_mrca2(self):
n = 1000
k = 6
step = 10
x = list(frange(0, 5000, step))
y = [coal.prob_mrca(i, k, n) * step for i in x]
y2 = cumsum(y)
y3 = [coal.cdf_mrca(t, k, n) for t in x]
y4 = [coal.prob_coal_counts(k, 1, t, n) for t in x]
fequals(y2, y3, eabs=.01)
fequals(y3, y4)
def test_cdf_coal_bounded(self):
n = 1000
k = 4
t = 500
step = .1
x = list(frange(0, 500, step))
y = [coal.prob_bounded_coal(i, k, n, t) * step for i in x]
y2 = cumsum(y)
y3 = [coal.cdf_bounded_coal(i, k, n, t) for i in x]
p = plot(x, y2, style="lines")
p.plot(x, y3, style="lines")
p.plot([0, 500], [1, 1], style="lines")
fequals(y2, y3, eabs=.01)
def optimize_model(self, gtree, stree, gene2species):
"""Optimizes the model"""
CostModel.optimize_model(self, gtree, stree, gene2species)
#=============================
# read sequences
if not self.align:
self.parser.error("--align must be specified")
self.align = fasta.read_fasta(self.align)
#=============================
# read SPIDIR parameters
if not self.params:
self.parser.error("--param must be specified")
self.params = spidir.read_params(self.params)
#=============================
# determine background base frequency
if self.bgfreq:
# use supplied frequency
vals = map(float, self.bgfreq.split(","))
if len(vals) != 4:
self.parser.error("invalid --bgfreq: %s" % self.bgfreq)
self.bgfreq = vals
else:
# compute frequency from alignment
self.bgfreq = alignlib.compute_bgfreq(self.align)
#=============================
# branch lengths
if self.kappa >= 0:
# use supplied kappa
self.kappa = self.kappa
else:
# compute kappa from alignment
# from spidir.find_ml_kapp_hky
minkappa = 0.4; maxkappa = 5.0; stepkappa = 0.1
maxlk = -util.INF
maxk = minkappa
for k in util.frange(minkappa, maxkappa, stepkappa):
l = spidir.find_ml_branch_lengths_hky(gtree, self.align, self.bgfreq, k, maxiter=1,
parsinit=(k == minkappa))
if l > maxlk:
maxlk = l
maxk = k
self.kappa = maxk
def make_color_legend(filename, colormap, start, end, step,
width=100, height=10):
from rasmus import util
s = svg.Svg(util.open_stream(filename, "w"))
s.beginSvg(width, height)
xscale = float(width) / (end + step - start)
for i in util.frange(start, end + step, step):
color = colormap.get(i)
s.rect((i-start) * xscale,
0,
step*xscale, height,
color, color)
s.endSvg()
def make_color_legend(filename, colormap, start, end, step,
width=100, height=10):
from rasmus import util
s = svg.Svg(util.open_stream(filename, "w"))
s.beginSvg(width, height)
xscale = float(width) / (end + step - start)
for i in util.frange(start, end + step, step):
color = colormap.get(i)
s.rect((i-start) * xscale,
0,
step*xscale, height,
color, color)
s.endSvg()
def test_prob_coal_cond_counts_simple(self):
# when we condition on b=1, it is the same as the bounded coal
# PDF.
# prob_coal_cond_counts is actually a more general version of
# prob_bounded_coal
outdir = 'test/tmp/test_coal/Coal_test_prob_coal_counys_simple/'
make_clean_dir(outdir)
a = 5
b = 1
t = 2000
n = 1000
p = Gnuplot()
p.enableOutput(False)
for x in frange(0, 2000, 10):
y = coal.prob_coal_cond_counts_simple(x, a, b, t, n)
y2 = coal.prob_bounded_coal(x, a, n, t)
self.assertAlmostEqual(y, y2)
def test_cdf_mrca(self):
outdir = 'test/tmp/test_coal/Coal_test_cdf_mrca/'
make_clean_dir(outdir)
n = 1000
k = 6
step = 10
x = list(frange(0, 5000, step))
y = [coal.prob_mrca(i, k, n) * step for i in x]
y2 = cumsum(y)
y3 = [coal.cdf_mrca(t, k, n) for t in x]
p = Gnuplot()
p.enableOutput(False)
p.plot(x, y2, style="lines")
p.plot(x, y3, style="lines")
p.enableOutput(True)
p.save(outdir + 'plot.png')
eq_sample_pdf(x, lambda t: coal.cdf_mrca(t, k, n), 40)
