def buildAlignBigTree(seqs, verbose=True, removetmp=True, options=""):
if len(seqs) < 2:
return seqs
# make input file for muscle
infilename = util.tempfile(".", "muscle-in", ".fa")
fasta.write_fasta(infilename, seqs)
# run muscle
outfilename = util.tempfile(".", "muscle-out", ".aln")
outfilename2 = util.tempfile(".", "muscle-out", ".tree")
cmd = "muscle -diags1 -sv -maxiters 1 " + options + " -in " + infilename + \
" -out " + outfilename + " -tree1 " + outfilename2
os.system(cmd)
# parse output
aln = fasta.read_fasta(outfilename)
tree = treelib.Tree()
tree.read_newick(outfilename2)
# cleanup tempfiles
if removetmp:
os.remove(infilename)
os.remove(outfilename)
os.remove(outfilename2)
return (aln, tree)
def muscle(seqs, verbose=True, removetmp=True, options=""):
if len(seqs) < 2:
return seqs
# make input file for muscle
infilename = util.tempfile(".", "muscle-in", ".fa")
fasta.write_fasta(infilename, seqs)
if not verbose:
options += " -quiet "
# run muscle
outfilename = util.tempfile(".", "muscle-out", ".aln")
cmd = "muscle " + options + " -in " + infilename + \
" -out " + outfilename
os.system(cmd)
# parse output
aln = fasta.read_fasta(outfilename)
# cleanup tempfiles
if removetmp:
os.remove(infilename)
os.remove(outfilename)
return aln
def buildAlignBigTree(seqs, verbose=True, removetmp=True, options=""):
if len(seqs) < 2:
return seqs
# make input file for muscle
infilename = util.tempfile(".", "muscle-in", ".fa")
fasta.write_fasta(infilename, seqs)
# run muscle
outfilename = util.tempfile(".", "muscle-out", ".aln")
outfilename2 = util.tempfile(".", "muscle-out", ".tree")
cmd = "muscle -diags1 -sv -maxiters 1 " + options + " -in " + infilename + \
" -out " + outfilename + " -tree1 " + outfilename2
os.system(cmd)
# parse output
aln = fasta.read_fasta(outfilename)
tree = treelib.Tree()
tree.read_newick(outfilename2)
# cleanup tempfiles
if removetmp:
os.remove(infilename)
os.remove(outfilename)
os.remove(outfilename2)
return (aln, tree)
def clustalw(seqs, verbose=True, removetmp=True, options=""):
"""Align sequences 'seqs' with clustalw"""
if len(seqs) < 2:
return seqs
# make input file for clustalw
infilename = util.tempfile(".", "clustalw-in", ".fa")
fasta.writeFasta(infilename, seqs)
# run clustalw
outfilename = util.tempfile(".", "clustalw-out", ".aln")
cmd = "clustalw " + options + " -quicktree -infile=" + infilename + \
" -outfile=" + outfilename
if not verbose:
cmd += " > /dev/null"
os.system(cmd)
# parse output
aln = readClustalwAlign(outfilename)
# cleanup tempfiles
if removetmp:
os.remove(infilename)
os.remove(outfilename)
os.remove(infilename.replace(".fa", ".dnd"))
# convert output
return aln
def bionj(aln=None, labels=None, distmat=None, seqtype="pep", verbose=True):
# make temp files
distfile = util.tempfile(".", "bionj-in", ".dist")
treefile = util.tempfile(".", "bionj-out", ".tree")
# find distances and then NJ tree
if distmat is not None:
phylip.write_dist_matrix(distmat, out=distfile)
if labels is None:
labels = aln.keys()
else:
if seqtype == "pep":
labels = phylip.protdist(aln, distfile, verbose=verbose)
else:
labels = phylip.dnadist(aln, distfile, verbose=verbose)
os.system("echo -n '%s\n%s' | bionj > /dev/null" % (distfile, treefile))
tree = treelib.read_tree(treefile)
phylip.rename_tree_with_names(tree, labels)
# clean up
os.remove(distfile)
os.remove(treefile)
return tree
def muscle(seqs, verbose=True, removetmp=True, options=""):
if len(seqs) < 2:
return seqs
# make input file for muscle
infilename = util.tempfile(".", "muscle-in", ".fa")
fasta.write_fasta(infilename, seqs)
if not verbose:
options += " -quiet "
# run muscle
outfilename = util.tempfile(".", "muscle-out", ".aln")
cmd = "muscle " + options + " -in " + infilename + \
" -out " + outfilename
os.system(cmd)
# parse output
aln = fasta.read_fasta(outfilename)
# cleanup tempfiles
if removetmp:
os.remove(infilename)
os.remove(outfilename)
return aln
def clustalwProfiles(aln1, aln2, verbose=True, removetmp=True, options=""):
# make input file for clustalw
infilename1 = util.tempfile(".", "clustalw-in", ".fa")
infilename2 = util.tempfile(".", "clustalw-in", ".fa")
fasta.writeFasta(infilename1, aln1)
fasta.writeFasta(infilename2, aln2)
# run clustalw
outfilename = util.tempfile(".", "clustalw-out", ".aln")
cmd = "clustalw " + options + " -quicktree -profile1=" + infilename1 + \
" -profile2=" + infilename2 + " -outfile=" + outfilename
if not verbose:
cmd += " > /dev/null"
os.system(cmd)
# parse output
aln = readClustalwAlign(outfilename)
# cleanup tempfiles
if removetmp:
os.remove(infilename1)
os.remove(infilename2)
os.remove(outfilename)
try:
os.remove(infilename1.replace(".fa", ".dnd"))
except:
pass
try:
os.remove(infilename2.replace(".fa", ".dnd"))
except:
pass
# convert output
keys = aln.keys()
return aln
def bl2seq(seq1, seq2, program="blastp", options="", name1="seq1", name2="seq2"):
"""
Performs Blast between two sequences 'seq1' and 'seq2'.
Returns a single Blast hit line or None if no hits found.
"""
# create temp files for sequences
file1 = util.tempfile(".", "blastp", ".fasta")
file2 = util.tempfile(".", "blastp", ".fasta")
fasta.write_fasta_ordered(file1, [name1], [seq1])
fasta.write_fasta_ordered(file2, [name2], [seq2])
# execute blast
pipe = os.popen("bl2seq -p %s -i %s -j %s -D 1 %s" % (program, file1, file2, options))
# parse hit
hit = None
for line in pipe:
if line[0] == "#":
continue
hit = line.rstrip().split("\t")
break
# remove temp files
os.remove(file1)
os.remove(file2)
return hit
def clustalwProfiles(aln1, aln2, verbose=True, removetmp=True, options=""):
# make input file for clustalw
infilename1 = util.tempfile(".", "clustalw-in", ".fa")
infilename2 = util.tempfile(".", "clustalw-in", ".fa")
fasta.writeFasta(infilename1, aln1)
fasta.writeFasta(infilename2, aln2)
# run clustalw
outfilename = util.tempfile(".", "clustalw-out", ".aln")
cmd = "clustalw " + options + " -quicktree -profile1=" + infilename1 + \
" -profile2=" + infilename2 + " -outfile=" + outfilename
if not verbose:
cmd += " > /dev/null"
os.system(cmd)
# parse output
aln = readClustalwAlign(outfilename)
# cleanup tempfiles
if removetmp:
os.remove(infilename1)
os.remove(infilename2)
os.remove(outfilename)
try:
os.remove(infilename1.replace(".fa", ".dnd"))
except:
pass
try:
os.remove(infilename2.replace(".fa", ".dnd"))
except:
pass
# convert output
keys = aln.keys()
return aln
def clustalw(seqs, verbose=True, removetmp=True, options=""):
"""Align sequences 'seqs' with clustalw"""
if len(seqs) < 2:
return seqs
# make input file for clustalw
infilename = util.tempfile(".", "clustalw-in", ".fa")
fasta.writeFasta(infilename, seqs)
# run clustalw
outfilename = util.tempfile(".", "clustalw-out", ".aln")
cmd = "clustalw " + options + " -quicktree -infile=" + infilename + \
" -outfile=" + outfilename
if not verbose:
cmd += " > /dev/null"
os.system(cmd)
# parse output
aln = readClustalwAlign(outfilename)
# cleanup tempfiles
if removetmp:
os.remove(infilename)
os.remove(outfilename)
os.remove(infilename.replace(".fa", ".dnd"))
# convert output
return aln
def bl2seq(seq1, seq2, program="blastp", options="",
name1="seq1", name2="seq2"):
"""
Performs Blast between two sequences 'seq1' and 'seq2'.
Returns a single Blast hit line or None if no hits found.
"""
# create temp files for sequences
file1 = util.tempfile(".", "blastp", ".fasta")
file2 = util.tempfile(".", "blastp", ".fasta")
fasta.write_fasta_ordered(file1, [name1], [seq1])
fasta.write_fasta_ordered(file2, [name2], [seq2])
# execute blast
pipe = os.popen("bl2seq -p %s -i %s -j %s -D 1 %s" %
(program, file1, file2, options))
# parse hit
hit = None
for line in pipe:
if line[0] == "#":
continue
hit = line.rstrip().split("\t")
break
# remove temp files
os.remove(file1)
os.remove(file2)
return hit
def create_temp_dir(prefix="tmpphylip_"):
"""Create a temporary directory for executing PHYLIP"""
directory = os.path.split(util.tempfile(".", prefix, ""))[1]
os.mkdir(directory)
os.chdir(directory)
return directory
def processFunc():
# remove old query tempfile if one exists
if closure["oldtmp"] != None:
os.remove(closure["oldtmp"])
elapse = util.toc()
closure["time"] += elapse
util.log("blasted %d of %d sequences (%.1f%%), elapse %.0f m, left %.0f m" % (
closure["index"], len(seqs.keys()),
100 * float(closure["index"]) / len(seqs.keys()),
closure["time"] / 60.0,
elapse / split * (len(seqs.keys()) - closure["index"]) / 60.0))
util.tic()
# find new subset of query sequences
i = closure["index"]
names = seqs.keys()[i:i+split]
# if no more sequences then quit
if len(names) == 0:
return False
# start blast
tmpfile = util.tempfile(".", "blastp", ".fasta")
seqs.write(tmpfile, names = names)
pipe = os.popen("blastall -p %s -d %s -i %s -m 8 -e .1 %s" % \
(prog, databaseFile, tmpfile, options))
# update variables
closure["oldtmp"] = tmpfile
closure["index"] = i + split
return pipe
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 showtab(tab, name='table'):
"""Show a table in a new xterm"""
name = name.replace("'", "")
tmp = util.tempfile(".", "tmp", ".tab")
tab.write_pretty(file(tmp, "w"))
os.system("(xterm -T '%s' -n '%s' -e less -S %s; rm %s) &" %
(name, name, tmp, tmp))
def gfit(self, func, eqn, params, list1, list2=[], list3=[], ** options):
"""
all syntax should be valid GNUPLOT syntax
func - a string of the function call i.e. "f(x)"
eqn - a string of a GNUPLOT equation "a*x**b"
params - a dictionary of parameters in eqn and their initial values
ex: {"a": 1, "b": 3}
"""
from rasmus import util
self.set(** options)
print len(list1), len(list2), len(list3)
if not self.enable:
raise Exception("must be output must be enabled for fitting")
list1, list2, list3 = self.prepareData(list1, list2, list3)
# add data to graph
self.data.append(self.Plot(list1, list2, list3, copy.copy(self.options)))
# perform fitting
self.stream = os.popen("gnuplot", "w")
print >>self.stream, "%s = %s" % (func, eqn)
for param, value in params.items():
print >>self.stream, "%s = %f" % (param, value)
print >>self.stream, "fit %s '-' via %s" % \
(func, ",".join(params.keys()))
self.outputData(list1, list2, list3)
# save and read parameters
outfile = util.tempfile(".", "plot", ".txt")
print >>self.stream, "save var '%s'" % outfile
print >>self.stream, "print 'done'"
self.stream.flush()
# wait for variable file
while not os.path.isfile(outfile): pass
params = self.readParams(outfile)
os.remove(outfile)
# build eqn for plotting
paramlist = ""
for param, value in params.items():
paramlist += "%s = %s, " % (param, value)
self.options["eqn"] = paramlist + "%s = %s, %s" % \
(func, eqn, func)
self.options["style"] = "lines"
# add fitted eqn to graph
self.data.append(self.Plot([], [], [], copy.copy(self.options)))
self.replot()
def wait(self):
"""Wait until all commands are known to be excuted"""
from rasmus import util
tmpfile = util.tempfile(".", "gnuplot", ".ps")
print >>self.stream, "set output '%s'" % tmpfile
print >>self.stream, "set terminal postscript color"
print >>self.stream, "plot '-'\n0 0\ne\n"
self.stream.flush()
while not os.path.isfile(tmpfile): pass
os.remove(tmpfile)
def buildTree(seqs, verbose=True, removetmp=True, options=""):
# make input file for muscle
infilename = util.tempfile(".", "muscle-in", ".fa")
fasta.write_fasta(infilename, seqs)
# run muscle
outfilename = util.tempfile(".", "muscle-out", ".tree")
cmd = "muscle " + options + " -in " + infilename + \
" -cluster -tree1 " + outfilename
if not verbose:
cmd += " 2>/dev/null"
os.system(cmd)
tree = treelib.Tree()
tree.read_newick(outfilename)
if removetmp:
os.remove(infilename)
os.remove(outfilename)
return tree
def setTerminal(self, filename = "", format="x11"):
if not self.enable:
return
from rasmus import util
# auto detect format from filename
if filename != "":
print >>self.stream, "set output \"%s\"" % filename
# determine format
if filename.endswith(".ps"):
format = "ps"
if filename.endswith(".pdf"):
format = "pdf"
if filename.endswith(".gif"):
format = "gif"
if filename.endswith(".png"):
format = "png"
if filename.endswith(".jpg"):
format = "jpg"
else:
tmpfile = util.tempfile(".", "gnuplot", ".ps")
print >>self.stream, "set output \"%s\"" % tmpfile
return tmpfile
# set terminal format
if format == "ps":
print >>self.stream, "set terminal postscript color"
elif format == "pdf":
print >>self.stream, "set terminal pdf"
elif format == "gif":
print >>self.stream, "set terminal gif"
elif format == "jpg":
print >>self.stream, "set terminal jpeg"
else:
print >>self.stream, "set terminal %s" % format
def buildTree(seqs, verbose=True, removetmp=True, options=""):
# make input file for clustalw
infilename = util.tempfile(".", "clustalw-in", ".fa")
fasta.writeFasta(infilename, seqs)
# run clustalw
outfilename = infilename.replace(".fa", ".ph")
cmd = "clustalw " + options + " -tree -infile=" + infilename + \
" -outfile=" + outfilename
if not verbose:
cmd += " > /dev/null"
os.system(cmd)
# parse output
tree = treelib.Tree()
tree.read_newick(outfilename)
# cleanup tempfiles
if removetmp:
os.remove(infilename)
os.remove(outfilename)
return tree
def heatmap(matrix,
width=20,
height=20,
colormap=None,
filename=None,
rlabels=None,
clabels=None,
display=True,
xdir=1,
ydir=1,
xmargin=0,
ymargin=0,
labelPadding=2,
labelSpacing=4,
mincutoff=None,
maxcutoff=None,
showVals=False,
formatVals=str,
valColor=black,
clabelsAngle=270,
clabelsPadding=None,
rlabelsAngle=0,
rlabelsPadding=None,
colors=None,
strokeColors=None,
valAnchor="start",
close=True):
from rasmus import util
if display and (not close):
raise Exception("must close file if display is used")
# determine filename
if filename is None:
filename = util.tempfile(".", "heatmap", ".svg")
temp = True
else:
temp = False
# determine colormap
if colors is None:
if colormap is None:
colormap = rainbowColorMap(util.flatten(matrix))
# determine matrix size and orientation
nrows = len(matrix)
ncols = len(matrix[0])
if xdir == 1:
xstart = xmargin
ranchor = "end"
coffset = width
elif xdir == -1:
xstart = xmargin + ncols * width
ranchor = "start"
coffset = 0
else:
raise Exception("xdir must be 1 or -1")
if ydir == 1:
ystart = ymargin
roffset = height
canchor = "start"
elif ydir == -1:
ystart = ymargin + nrows * width
roffset = 0
canchor = "end"
else:
raise Exception("ydir must be 1 or -1")
# begin svg
infile = util.open_stream(filename, "w")
s = svg.Svg(infile)
s.beginSvg(ncols * width + 2 * xmargin, nrows * height + 2 * ymargin)
# draw matrix
for i in xrange(nrows):
for j in xrange(ncols):
if mincutoff and matrix[i][j] < mincutoff:
continue
if maxcutoff and matrix[i][j] > maxcutoff:
continue
if colors:
color = colors[i][j]
else:
color = colormap.get(matrix[i][j])
if strokeColors:
strokeColor = strokeColors[i][j]
else:
strokeColor = color
s.rect(xstart + xdir * j * width, ystart + ydir * i * height,
xdir * width, ydir * height, strokeColor, color)
# draw values
if showVals:
# find text size
fontwidth = 7 / 11.0
textsize = []
for i in xrange(nrows):
for j in xrange(ncols):
if mincutoff and matrix[i][j] < mincutoff:
continue
if maxcutoff and matrix[i][j] > maxcutoff:
continue
strval = formatVals(matrix[i][j])
if len(strval) > 0:
textsize.append(
min(height, width / (float(len(strval)) * fontwidth)))
textsize = min(textsize)
if valAnchor == "start":
xoffset = 0
elif valAnchor == "middle":
xoffset = 0.5
elif valAnchor == "end":
xoffset = 1
else:
raise Exception("anchor not supported: %s" % valAnchor)
yoffset = int(ydir == -1)
for i in xrange(nrows):
for j in xrange(ncols):
if mincutoff and matrix[i][j] < mincutoff:
continue
if maxcutoff and matrix[i][j] > maxcutoff:
continue
strval = formatVals(matrix[i][j])
s.text(strval,
xstart + xdir * (j + xoffset) * width,
ystart + ydir * (i + yoffset) * height + height / 2.0 +
textsize / 2.0,
textsize,
fillColor=valColor,
anchor=valAnchor)
# draw labels
if rlabels is not None:
assert len(rlabels) == nrows, \
"number of row labels does not equal number of rows"
if rlabelsPadding is None:
rlabelsPadding = labelPadding
for i in xrange(nrows):
x = xstart - xdir * rlabelsPadding
y = ystart + roffset + ydir * i * height - labelSpacing / 2.
s.text(rlabels[i],
x,
y,
height - labelSpacing,
anchor=ranchor,
angle=rlabelsAngle)
if clabels is not None:
assert len(clabels) == ncols, \
"number of col labels does not equal number of cols"
if clabelsPadding is None:
clabelsPadding = labelPadding
for j in xrange(ncols):
x = xstart + coffset + xdir * j * width - labelSpacing / 2.
y = ystart - ydir * clabelsPadding
s.text(clabels[j],
x,
y,
width - labelSpacing,
anchor=canchor,
angle=clabelsAngle)
# end svg
if close:
s.endSvg()
s.close()
# display matrix
if display:
#if temp:
os.system("display %s" % filename)
#else:
# os.spawnl(os.P_NOWAIT, "display", "display", filename)
# clean up temp files
if temp:
os.remove(filename)
return s
