def my_layout(node):
if node.is_leaf() and any(
[node.name == x.split(';')[-1] for x in selection['Taxonomy_UniEuk']]):
F = ete3.TextFace(node.name, tight_text=True, fgcolor='red')
else:
F = ete3.TextFace(node.name, tight_text=True)
ete3.add_face_to_node(F, node, column=0, position="branch-right")
def buildTreeDepthFirst(expr,parentList):
if expr[0]==';':
leafNodes=buildTreeDepthFirst(expr[1],parentList)
treeNodes=buildTreeDepthFirst(expr[2],leafNodes)
return treeNodes
elif expr[0]=='if-then:':
leafNodes1=buildTreeDepthFirst(expr[1],parentList)
successLeafs=buildTreeDepthFirst(expr[2],leafNodes1)
failActionNodes=[]
for parent in parentList:
c=parent.add_child(name='Failed Action')
c.add_face(ete3.TextFace('Failed Action'),column=0)
failActionNodes.append(c)
failLeafs=buildTreeDepthFirst(expr[3],failActionNodes)
return successLeafs+failLeafs
#elif expr[0].find("Loop-for-") != -1:
# line=expr[0][9:]
# pos=line.find('-')
# loopCount=line[:pos]
# loopCount=int(loopCount)
# tree1=buildTreeDepthFirst(expr[1],parent)
# treeList=[tree1 for i in xrange(loopCount)]
# j=loopCount-2
# currentTree=treeList[loopCount-1]
# while j > -1:
# currentTree=appendTrees(treeList[j],currentTree)
# j=j-1
# return currentTree
else:
resultList=[]
for parent in parentList:
c=parent.add_child(name=expr)
c.add_face(ete3.TextFace(expr),column=0)
resultList.append(c)
return resultList
def displaySVsOnPhylogenies(pid, t, SVs):
if pid in SVs:
for SV in SVs[pid]:
SVname = SV
text_face = ete3.TextFace(SVname, fsize=vertical_margins[pid] - 5)
branches = getBranchOrBranchesFor(SVs[pid][SV], t, set(), set(),
set())
if len(branches) > 1:
text_face = ete3.TextFace(SVname,
fsize=vertical_margins[pid] - 5,
fgcolor="orangered")
for branch in branches:
branch.add_face(text_face, column=1, position="branch-top")
#Collect data for statistics
samples = SVs[pid][SV]
ntips = 0
for tip in t:
ntips += 1
if not displayCombo and len(samples) > 1 and len(samples) < ntips:
SV_out.write(str(pid))
SV_out.write("\t" + SV)
if len(branches) > 1:
SV_out.write("\tMiss")
else:
SV_out.write("\tMatch")
SV_out.write("\n")
else:
print("No", pid, "in SVs")
def add_legend(tstyle, varname, all_vals, smap, info, start_column, add_missing=False, add_sign=None, reverse_log=False, n_entries=5, fsize=4, no_opacity=False): # NOTE very similar to add_smap_legend() in plot_2d_scatter() in python/lbplotting.py
if len(all_vals) == 0:
return
assert add_sign in [None, '-', '+']
tstyle.legend.add_face(ete3.TextFace(' %s ' % varname, fsize=fsize), column=start_column)
min_val, max_val = get_scale_min(args.lb_metric, all_vals), max(all_vals)
if min_val == max_val:
return
max_diff = (max_val - min_val) / float(n_entries - 1)
val_list = list(numpy.arange(min_val, max_val + utils.eps, max_diff)) # first value is exactly <min_val>, last value is exactly <max_val> (eps is to keep it from missing the last one)
# if add_sign is not None and add_sign == '-': # for negative changes, we have the cmap using abs() and want to legend order to correspond
# val_list = reversed(val_list) # arg, this breaks something deep in the legend maker, not sure what
key_list = [None for _ in val_list]
if add_missing:
val_list += [None]
key_list += ['missing!'] # doesn't matter what the last one is as long as it isn't in <affyfo>
for val, key in zip(val_list, key_list):
tstyle.legend.add_face(ete3.TextFace('', fsize=fsize), column=start_column)
if smap is None:
sz = get_size(min_val, max_val, val)
rface = ete3.RectFace(sz, sz, bgcolor=plotting.getgrey(), fgcolor=None)
else:
rface = ete3.RectFace(6, 6, bgcolor=plotting.get_smap_color(smap, info, key=key, val=val), fgcolor=None)
if not no_opacity:
rface.opacity = opacity
tstyle.legend.add_face(rface, column=start_column + 1)
fstr = '%.1f' if args.lb_metric == 'cons-dist-aa' else '%.2f'
tstyle.legend.add_face(ete3.TextFace(((' %s'+fstr) % (add_sign if add_sign is not None else '', math.exp(val) if reverse_log else val)) if key is None else ' missing', fsize=fsize), column=start_column + 2)
def displayDriversOnPhylogenies(pid, t, drivers, deletions):
if pid in drivers:
for gene in drivers[pid]:
append = ""
if len(drivers[pid][gene]) > 1:
append = "_a"
for posalt in drivers[pid][gene]:
text_face = ete3.TextFace(gene + append,
fsize=vertical_margins[pid] - 5)
incongruous_samples = set()
alt_samples = set()
dels = set()
if pid in deletions and gene in deletions[pid] and posalt[
0] in deletions[pid][gene]:
dels = deletions[pid][gene][posalt[0]]
branches = getBranchOrBranchesFor(drivers[pid][gene][posalt],
t, incongruous_samples,
alt_samples, dels)
if len(branches) > 1:
text_face = ete3.TextFace(gene + append,
fsize=vertical_margins[pid] - 5,
fgcolor="orangered")
for branch in branches:
branch.add_face(text_face, column=1, position="branch-top")
if checkDelete:
for sample in alt_samples:
del_look_file.write(str(pid))
del_look_file.write("\t" + sample)
del_look_file.write("\t" + gene2chrom[gene])
del_look_file.write("\t" + posalt[0])
del_look_file.write("\t" + gene)
del_look_file.write("\t" +
str(sample in incongruous_samples))
del_look_file.write("\n")
if displayDeletions and len(dels) > 0 and len(
drivers[pid][gene][posalt]) > 0:
text_face = ete3.TextFace(u'Δ ' + gene + append,
fsize=vertical_margins[pid] - 5)
#print(str(dels))
branches = getBranchOrBranchesFor(dels, t, set(), set(),
set())
for branch in branches:
branch.add_face(text_face,
column=1,
position="branch-top")
if append == "_a":
append = "_b"
elif append == "_b":
append = "_c"
elif append == "_c":
append = "_d"
elif append == "_d":
append = "_e"
else:
print("No", pid, "in drivers")
def buildTreeDepthFirst(tranDict, t, nextState):
if nextState in tranDict:
possibleTrans = tranDict[nextState]
for tran in possibleTrans:
if tran.action == "finishedAction":
buildTreeDepthFirst(tranDict, t, tran.endState)
else:
c = t.add_child(name=tran.action)
c.dist = tran.percentage
if float(tran.percentage) > 0.5:
c.add_face(ete3.TextFace(tran.action), column=0)
else:
c.add_face(ete3.TextFace("action failed"), column=0)
buildTreeDepthFirst(tranDict, c, tran.endState)
def mylayout(node):
node.set_style(ns)
if not node.is_leaf():
ete3.add_face_to_node(ete3.TextFace(node.name),
node,
column=0,
position='branch-top')
ete3.add_face_to_node(ete3.TextFace('\n'.join(
getattr(node, 'info', []))),
node,
column=0,
position='branch-bottom')
if node.support <= 0.5:
node.set_style(dashed_branch)
add_img(node)
def layout(node):
""" layout for ete2 tree plotting fig """
if node.is_leaf():
nameF = ete3.TextFace(node.name, tight_text=False,
fgcolor="#262626", fsize=8)
ete3.add_face_to_node(nameF, node, column=0, position="aligned")
node.img_style["size"] = 0
else:
if not node.is_root():
node.img_style["size"] = 0
node.img_style["shape"] = 'square'
node.img_style["fgcolor"] = "#262626"
if "quartets_total" in node.features:
ete3.add_face_to_node(ete2.PieChartFace(
percents=[float(node.quartets_sampled_prop),
100-float(node.quartets_sampled_prop)],
width=15, height=15, colors=PIECOLORS),
node, column=0, position="float-behind")
if "bootstrap" in node.features:
ete3.add_face_to_node(ete2.AttrFace(
"bootstrap", fsize=7, fgcolor='red'),
node, column=0, position="branch-top")
else:
node.img_style["size"] = 0
if node.is_root():
node.img_style["size"] = 0
node.img_style["shape"] = 'square'
node.img_style["fgcolor"] = "#262626"
def _align_leaf_name(tree):
attr_face = ete3.AttrFace('name')
text_face = ete3.TextFace(' ')
for leaf in tree:
leaf.add_face(text_face, column=0, position='aligned')
leaf.add_face(attr_face, column=1, position='aligned')
return 0
def evolve_tree(newick,
dup_mut=0.00043,
del_mut=0.0,
dup_rev=0.00086,
del_rev=0.0):
t = copy.deepcopy(newick)
mutstyle = ete3.NodeStyle()
mutstyle['fgcolor'] = 'red'
gen_per_snp = 1 / (
25 * .0076
) #Generations per unit of branch length/SNP; .0076 is SNPs per year, 25 is years per generation
for node in t.traverse():
if node.is_root():
node.add_feature('cnv_state', 'reference')
continue
node.add_feature('cnv_state', node.up.cnv_state)
orig_state = node.cnv_state
for generation in xrange(
int(node.dist * gen_per_snp)
): #maybe change this to a power later--would require editing evolve_node function
evolve_node(node,
dup_mut=dup_mut,
del_mut=del_mut,
dup_rev=dup_rev,
del_rev=del_rev)
if node.cnv_state != orig_state:
node.set_style(mutstyle)
if node.is_leaf() and node.cnv_state != 'reference':
node.add_face(ete3.TextFace(' %s' % (node.cnv_state[:3])), 1,
'branch-right')
return t
def my_layout(node):
name = getattr(node, attribute_name)
try:
kmer_full = locale.format("%d", int(node.kmers_full), grouping=True),
except AttributeError:
kmer_full = None
try:
kmer_reduced = locale.format("%d", int(node.kmers_reduced), grouping=True)
except AttributeError:
kmer_reduced = None
if kmer_full is None:
if kmer_reduced is None:
t = name
else:
t = "{} [red. {}]".format(name, kmer_reduced)
else:
if kmer_reduced is None:
t = "{} [full {}]".format(name, kmer_full)
else:
t = "{} [full {} & red. {}]".format(name, kmer_full, kmer_reduced)
f = ete3.TextFace(t, tight_text=True)
ete3.add_face_to_node(f, node, column=0, position="branch-right")
def draw_tree(plotdir,
plotname,
treestr,
gl_sets,
all_genes,
gene_categories,
ref_label=None,
arc_start=None,
arc_span=None):
etree = ete3.ClusterTree(treestr)
node_names = set() # make sure we get out all the genes we put in
for node in etree.traverse():
if set_distance_to_zero(node):
node.dist = 0. if ref_label is not None else 1e-9 # data crashes sometimes with float division by zero if you set it to 0., but simulation sometimes gets screwed up for some other reason (that I don't understand) if it's 1e-9
# node.dist = 1.
status = getstatus(gene_categories, node, ref_label=ref_label)
set_node_style(node, status, len(gl_sets), ref_label=ref_label)
if node.is_leaf():
node_names.add(node.name)
if len(set(all_genes) - node_names) > 0:
raise Exception('missing genes from final tree: %s' %
' '.join(node_names))
if ref_label is None: # have to do it in a separate loop so it doesn't screw up the distance setting
for node in [n for n in etree.traverse()
if n.is_leaf()]: # yeah I'm sure there's a fcn for that
node.name = shorten_name(node.name)
tstyle = ete3.TreeStyle()
tstyle.show_scale = False
if not args.leaf_names:
tstyle.show_leaf_name = False
# tstyle.mode = 'c'
# if arc_start is not None:
# tstyle.arc_start = arc_start
# if arc_span is not None:
# tstyle.arc_span = arc_span
write_legend(plotdir)
if args.title is not None:
fsize = 13
tstyle.title.add_face(ete3.TextFace(args.title, fsize=fsize,
bold=True),
column=0)
if args.title_color is not None:
# tstyle.title.add_face(ete3.CircleFace(fsize, scolors[args.title]), column=1)
tcol = scolors[
args.
title_color] if args.title_color in scolors else args.title_color
rect_width = 3 if len(args.title) < 12 else 2
tstyle.title.add_face(ete3.RectFace(width=rect_width * fsize,
height=fsize,
bgcolor=tcol,
fgcolor=None),
column=1)
suffix = '.svg'
imagefname = plotdir + '/' + plotname + suffix
print ' %s' % imagefname
etree.render(imagefname, tree_style=tstyle)
def generate_heatmap(self, ete3_tree, *args):
for node in ete3_tree:
if node.is_leaf():
taxon = node.name
tc_index = 0
for args_index in range(0, len(args), 2):
taxon_to_classification = args[args_index]
class_to_color = args[args_index + 1]
try:
classification = taxon_to_classification[taxon]
color = class_to_color[classification]
except:
classification = ''
color = 'white'
class_face = ete3.TextFace(classification,
ftype='Arial',
fsize=2,
fgcolor='white')
class_face.background.color = color
class_face.hz_align = 1
class_face.vt_align = 1
class_face.margin_left = 5
class_face.margin_right = 5
node.add_face(class_face, tc_index, "aligned")
tc_index += 1
return ete3_tree
def fam_size_layout(node):
try:
node_color = self.graphics_options[node.event]
except:
node_color = self.graphics_options['=']
relative_size = 8 * (node.fam_size / max(family.fam_sizes))
cf = ete3.CircleFace(
radius=relative_size * self.graphics_options['scale'],
color=node_color,
style='circle'
)
cf.opacity = self.graphics_options['opacity']
node.add_face(cf, column=10, position='float')
# add the family size number and asterisks to the figure
famsize_str = '{}{}\n'.format(
get_pvalue_asterisks(node),
node.fam_size,
)
tf = ete3.TextFace(famsize_str, fsize=4, fgcolor=node_color)
node.add_face(tf, column=1, position='float')
# remove the silly default blue node dot
nstyle = ete3.NodeStyle()
nstyle['size'] = 0
node.set_style(nstyle)
return
def style_subject_node(color="Black"):
"""Specify the appearance of Subject nodes."""
face = ete3.TextFace(node.name,
fsize=SUBJECT_NODE_FONT_SIZE,
fgcolor=color)
set_face_margin(face)
node.add_face(face, column=0, position="branch-right")
def _show_inner_name(tree, tree_format):
if tree_format != 0:
return 0
for node in tree.traverse():
if (not node.is_leaf()) and (not node.is_root()):
node_name_face = ete3.TextFace(node.name)
node.add_face(node_name_face, position='branch-bottom', column=0)
return 0
def summary_tree(self):
'''
show a tree that visualizes the total number of expansions and
contractions across the whole phylogeny for allgene families.
'''
def fam_size_piechart_layout(node):
'''
the PieChart layout function, defined in local scope so it
can access class attributes (graphics options).
'''
if not node.is_root():
if self.count_all_expansions:
n_exp = node.expansion
n_con = node.decrease
else:
n_exp = node.sig_expansions
n_con = node.sig_expansions
if hasattr(self, 'lambda_colors'):
circle_color = self.lambda_colors[node.lambda_group]
else:
circle_color = "blue"
# add a text that shows expansions & contractions, e.g. +10/-20
exp_cnt_txt = ete3.TextFace(
'+{} -{}\n'.format(int(n_exp), int(n_con)), fsize=6,
fgcolor=circle_color
)
pos = 'aligned' if node.is_leaf() else 'float'
# add a circle that shows the average expansion
ete3.faces.add_face_to_node(exp_cnt_txt, node, 1, position=pos)
# add average expansion info:
scale_factor = 1 + node.avg_expansion
avg_exp = '{:+}'.format(round(node.avg_expansion, 2))
circle = ete3.CircleFace(radius=9 * scale_factor,
color=circle_color,
label={'text': avg_exp, 'color': 'white',
'fontsize': 3 + (2.25*scale_factor)})
circle.opacity = self.graphics_options['opacity']
ete3.faces.add_face_to_node(circle, node, 2, position='float')
nstyle = ete3.NodeStyle()
nstyle['size'] = 0
node.set_style(nstyle)
return
t = self.tree
ts = ete3.TreeStyle()
header = 'Family expansions and contractions'
if hasattr(self, "lambdas"):
header += ('\nmin lambda: {} '
'(blue)\nmax lambda: {} (red)').format(
min(self.lambdas.values()), max(self.lambdas.values()))
ts.title.add_face(ete3.TextFace(header, fsize=8), column=0)
ts.scale = self.graphics_options['pixels_per_mya'] # pixels per million years
ts.layout_fn = fam_size_piechart_layout
self.show_or_dump_tree(tree_obj=t, tree_style=ts, fname='summary')
return
def branch_category_layout(node):
nstyle = ete3.NodeStyle()
nstyle['size'] = 0
nstyle["hz_line_width"] = nstyle["vt_line_width"] = 1
nstyle["hz_line_color"] = node.color
nstyle["vt_line_color"] = node.color
nlabel = node.name+'|'+str(node.numerical_label)
nlabelFace = ete3.TextFace(nlabel, fsize=4, fgcolor=node.color)
ete3.add_face_to_node(face=nlabelFace, node=node, column=1, aligned=False, position="branch-right")
node.set_style(nstyle)
def _node_layout(node):
global _circle_tested
node.img_style["size"] = 0
if node.is_leaf():
if node.name not in annotation:
print('Got unknown leaf "%s"' % (node.name))
return
size = annotation[node.name]
else:
children = COMMA.join(sorted([leaf.name for leaf in node]))
if children not in annotation:
print('Got unknown node ' + children)
return
size = annotation[children]
dimension = 4 * math.sqrt(size)
labeld = {
'text': "%d" % (size),
'color': 'white',
'font': 'Helvetica',
'size': 8
}
if size % 2:
clabel = dict(labeld)
if size == 35:
if not _circle_tested:
_circle_tested = True
clabel['text'] = 'Ly'
clabel['size'] = 12
thisFace = ete.CircleFace(dimension / 2,
"steelblue",
"circle",
label=clabel)
elif size == 43:
clabel['size'] = 6
del clabel['color']
thisFace = ete.CircleFace(dimension / 2,
"steelblue",
"sphere",
label=clabel)
else:
thisFace = ete.CircleFace(dimension / 2,
"steelblue",
"sphere",
label="%d" % (size))
else:
thisFace = ete.RectFace(dimension,
dimension,
'green',
'blue',
label=labeld)
thisFace.opacity = 0.7
ete.add_face_to_node(thisFace, node, column=0, position="float")
textF = ete.TextFace(str(size), fsize=12, fgcolor="steelblue")
ete.add_face_to_node(textF, node, column=0, position="aligned")
def build_tree(jsonNode, tr=ete3.Tree(), depth=0):
if type(jsonNode) is dict:
for key, value in jsonNode.items():
this_node = tr.add_child(name=str(jsonNode))
face = ete3.TextFace(key)
this_node.add_face(face, column=1, position="branch-top")
for child in value:
build_tree(child, this_node, depth + 1)
else:
tr.add_child(name=str(jsonNode))
return tr
def add_child(self,
attr,
val,
cond=None,
df=None,
target=None,
continuous=False):
if attr is None:
if not continuous:
distribution = round(len(df[df[target] == val]) / len(df),
2) * 100
msg = ' {}: {}% '.format(val, int(distribution))
else:
msg = ' {}: {} '.format(target, round(val, 2))
elif not continuous:
msg = ' {} = {} '.format(attr, val)
else:
msg = ' {} {} {} '.format(attr, '{}', round(val, 2))
if len(self.children) == 0:
ete3_child = self.ete3_node.add_child(name=msg)
else:
ete3_child = self.children[-1].ete3_node.add_sister()
if attr is not None and not continuous:
ete3_child.add_face(ete3.TextFace(msg),
column=0,
position='branch-right')
elif attr is not None and len(self.children) == 0:
ete3_child.add_face(ete3.TextFace(msg.format('<=')),
column=0,
position='branch-right')
elif attr is not None:
ete3_child.add_face(ete3.TextFace(msg.format('>')),
column=0,
position='branch-right')
node = Tree(attr, val, cond, ete3_node=ete3_child)
self.children.append(node)
return node
def draw_tree(plotdir, plotname, treestr, gl_sets, all_genes, gene_categories, ref_label=None, arc_start=None, arc_span=None):
etree = ete3.ClusterTree(treestr)
node_names = set() # make sure we get out all the genes we put in
for node in etree.traverse():
if set_distance_to_zero(node):
node.dist = 0. if ref_label is not None else 1e-9 # data crashes sometimes with float division by zero if you set it to 0., but simulation sometimes gets screwed up for some other reason (that I don't understand) if it's 1e-9
# node.dist = 1.
status = getstatus(gene_categories, node, ref_label=ref_label)
set_node_style(node, status, len(gl_sets), ref_label=ref_label)
if node.is_leaf():
node_names.add(node.name)
if len(set(all_genes) - node_names) > 0:
raise Exception('missing genes from final tree: %s' % ' '.join(node_names))
if args.param_dirs is not None:
countfo = OrderedDict()
for label, pdir in zip(args.glslabels, args.param_dirs): # it would be cleaner to do this somewhere else
if pdir == 'None': # not the best way to do this
continue
countfo[label] = utils.read_overall_gene_probs(pdir, normalize=True)[args.region]
for node in etree.traverse():
node.countstr = '%s' % ' '.join([('%.2f' % (100 * cfo[node.name])) if node.name in cfo else '-' for cfo in countfo.values()])
if ref_label is None: # have to do it in a separate loop so it doesn't screw up the distance setting
for node in [n for n in etree.traverse() if n.is_leaf()]: # yeah I'm sure there's a fcn for that
node.name = utils.shorten_gene_name(node.name)
tstyle = ete3.TreeStyle()
tstyle.show_scale = False
if len(args.glslabels) > 1:
write_legend(plotdir)
if args.title is not None:
fsize = 13
tstyle.title.add_face(ete3.TextFace(args.title, fsize=fsize, bold=True), column=0)
if args.title_color is not None:
# tstyle.title.add_face(ete3.CircleFace(fsize, scolors[args.title]), column=1)
tcol = scolors[args.title_color] if args.title_color in scolors else args.title_color
rect_width = 3 if len(args.title) < 12 else 2
tstyle.title.add_face(ete3.RectFace(width=rect_width*fsize, height=fsize, bgcolor=tcol, fgcolor=None), column=1)
suffix = '.svg'
imagefname = plotdir + '/' + plotname + suffix
print ' %s' % imagefname
etree.render(utils.insert_before_suffix('-leaf-names', imagefname), tree_style=tstyle)
tstyle.show_leaf_name = False
etree.render(imagefname, tree_style=tstyle)
# NOTE all the node names are screwed up after this, so you'll have to fix them if you add another step
if args.param_dirs is not None:
for node in etree.traverse():
node.name = node.countstr
tstyle.show_leaf_name = True
etree.render(utils.insert_before_suffix('-gene-counts', imagefname), tree_style=tstyle)
def branch_state_layout(node):
nstyle = ete3.NodeStyle()
nstyle['size'] = 0
nstyle["hz_line_width"] = nstyle["vt_line_width"] = 1
nstyle["hz_line_color"] = node.color
nstyle["vt_line_color"] = node.color
if node.is_leaf():
nlabel = str(node.state)+'|'+node.name
else:
nlabel = str(node.state)
nlabelFace = ete3.TextFace(nlabel, fsize=6, fgcolor=node.color)
ete3.add_face_to_node(face=nlabelFace, node=node, column=1, aligned=False, position="branch-right")
node.set_style(nstyle)
def _main():
print("Reading matrix from file")
root = None
output_file_name = None
if len(sys.argv) < 2:
print("Please supply at least one path to a distance matrix")
quit(1)
elif sys.argv[1] == "matrix": # Usage: python3 PhyloTree.py matrix OUTPUT_FILE_NAME [PATH_TO_INPUT_MATRICES . . .]
node_names_list = []
dist_mat_list = []
for mat in sys.argv[3:]:
(node_names, dist_mat) = Utility.read_distance_matrix(mat)
node_names_list.append(node_names)
dist_mat_list.append(dist_mat)
(node_names, dist_mat) = Utility.combine_distance_matrices(node_names_list, dist_mat_list, norm=False)
print("Writing combined matrix to {}".format(sys.argv[2]))
Utility.write_distance_matrix(dist_mat, node_names, sys.argv[2])
exit(0)
elif len(sys.argv) == 2: # Only one matrix provided
# Read the distance matrix in to memory
(node_names, dist_mat) = Utility.read_distance_matrix(sys.argv[1])
print("Creating tree with Neighbor-Joining from 1 matrix")
# Our neighbor joining algorithm creates rooted trees
# the root is created a node in the middle of the final edge added to the tree
root = create_nj_tree(node_names, dist_mat)
gene_name = sys.argv[1].split("/")[-1]
output_file_name = "{}_tree.png".format(gene_name)
else: # more than one matrix, average them then make tree
node_names_list = []
dist_mat_list = []
for mat in sys.argv[1:]:
(node_names, dist_mat) = Utility.read_distance_matrix(mat)
node_names_list.append(node_names)
dist_mat_list.append(dist_mat)
(node_names, dist_mat) = Utility.combine_distance_matrices(node_names_list, dist_mat_list, norm=False)
print("Creating tree with Neighbor-Joining from {} matrices".format(len(sys.argv[1:])))
root = create_nj_tree(node_names, dist_mat)
system_name = sys.argv[1].split("/")[-1]
output_file_name = "{}_combined_tree.png".format(system_name)
ts = ete3.TreeStyle()
ts.show_branch_length = True
ts.title.add_face(ete3.TextFace(output_file_name[0:-4], fsize=20), column=0)
nstyle = ete3.NodeStyle()
nstyle["shape"] = "sphere"
nstyle["size"] = 6
nstyle["fgcolor"] = "darkred"
for n in root.traverse():
n.set_style(nstyle)
root.show(tree_style=ts)
root.render(output_file_name, tree_style=ts, units="px", w=600)
def _build_node_text_face(node, color_node_labels, color, label_colors,
fontsize, rename_function):
if color_node_labels:
if label_colors is None:
node_color = color
elif type(label_colors) == dict:
node_color = label_colors.get(node.name, '#000000')
elif type(label_colors) in [list, tuple]:
node_color = color if node.name in label_colors else '#000000'
else:
node_color = '#000000'
else:
node_color = '#000000'
node_name = node.name if rename_function is None else rename_function(
node.name)
tf = ete3.TextFace(node_name, fsize=fontsize, fgcolor=node_color)
return tf
def process_tree(self, t):
# Add faces
if self.attribute:
values = set([getattr(l, self.attribute) for l in t.get_leaves()])
colours = get_colour_set(len(values))
colour_map = dict(zip(values, colours))
for l in t.iter_leaves():
mycolour = colour_map[getattr(l, self.attribute)]
if not self.dummy:
l.add_face(
ete3.CircleFace(radius=10,
color=mycolour,
style="sphere"), 0)
# Apply labels
if not self.dummy:
for l in t.iter_leaves():
l.add_face(ete3.TextFace(getattr(l, self.label)), 1)
# Plot or save
if self.output:
kw = {}
if self.height or self.width:
kw["h"] = self.height
kw["w"] = self.width
kw["units"] = self.units
kw["dpi"] = self.dpi
if self.multiple:
base, ext = os.path.splitext(self.output)
filename = base + ("_%06d" % (self.n + 1)) + ext
else:
filename = self.output
if not self.dummy:
t.render(filename, ultrametric, tree_style=self.ts, **kw)
else: # pragma: no cover
if not self.dummy:
t.show(ultrametric, tree_style=self.ts)
self.n += 1
if self.multiple:
return None
else:
raise StopIteration
def make_tree_vis(newick_tree, gene_name, file_name):
"""
Create visualizations of maximum-likelihood trees generated using ete
"""
tree = et.Tree(newick_tree)
#Customizations
nifty = et.TreeStyle()
nifty.show_branch_support = True
nifty.branch_vertical_margin = 15
nifty.scale = 10
#Leaf node customizations
style_L = et.NodeStyle()
style_L['shape'] = 'sphere'
style_L['size'] = 10
style_L['fgcolor'] = '#6BC245'
style_L['hz_line_type'] = 1
style_L['hz_line_color'] = '#cccccc'
#Internal node customizations
style_I = et.NodeStyle()
style_I['shape'] = 'sphere'
style_I['size'] = 10
style_I['fgcolor'] = '#5271FF'
style_I['hz_line_type'] = 1
style_I['hz_line_color'] = '#cccccc'
#Apply stylizations
for node in tree.traverse():
if node.is_leaf():
node.set_style(style_L)
else:
node.set_style(style_I)
#Annotate root of tree with gene name
gene_label = et.TextFace(gene_name, ftype="Helvetica", fsize=5,\
fgcolor='green')
tree.add_face(gene_label, column=0, position="branch-top")
tree.render(file_name + "_tree.png", tree_style=nifty)
def fam_size_piechart_layout(node):
'''
the PieChart layout function, defined in local scope so it
can access class attributes (graphics options).
'''
if not node.is_root():
if self.count_all_expansions:
n_exp = node.expansion
n_con = node.decrease
else:
n_exp = node.sig_expansions
n_con = node.sig_expansions
if hasattr(self, 'lambda_colors'):
circle_color = self.lambda_colors[node.lambda_group]
else:
circle_color = "blue"
# add a text that shows expansions & contractions, e.g. +10/-20
exp_cnt_txt = ete3.TextFace(
'+{} -{}\n'.format(int(n_exp), int(n_con)), fsize=6,
fgcolor=circle_color
)
pos = 'aligned' if node.is_leaf() else 'float'
# add a circle that shows the average expansion
ete3.faces.add_face_to_node(exp_cnt_txt, node, 1, position=pos)
# add average expansion info:
scale_factor = 1 + node.avg_expansion
avg_exp = '{:+}'.format(round(node.avg_expansion, 2))
circle = ete3.CircleFace(radius=9 * scale_factor,
color=circle_color,
label={'text': avg_exp, 'color': 'white',
'fontsize': 3 + (2.25*scale_factor)})
circle.opacity = self.graphics_options['opacity']
ete3.faces.add_face_to_node(circle, node, 2, position='float')
nstyle = ete3.NodeStyle()
nstyle['size'] = 0
node.set_style(nstyle)
return
def plot_tree(self, pdf_fn):
print("Plotting", pdf_fn, file=sys.stderr)
ts = ete3.TreeStyle()
ant = self.antibiotics.upper()
ts.title.add_face(ete3.TextFace("{}{}".format(18 * " ", ant),
fsize=60),
column=0)
ts.show_leaf_name = False
for n in self.tree.traverse():
taxid = n.name
nstyle = ete3.NodeStyle()
res_cat = self.res_dict[taxid]
nstyle["fgcolor"] = cat_to_color[res_cat]
nstyle["size"] = 15
n.set_style(nstyle)
if n.is_leaf():
N = ete3.AttrFace("name", fsize=20)
n.add_face(N, 0)
#add_face_to_node(N, n, 0, position="aligned")
#n.add_face_to_node(ete3.TextFace("XXX"))
self.tree.render(pdf_fn, tree_style=ts)
def set_meta_styles(args, etree, tstyle):
lbfo = args.metafo[args.lb_metric]
if args.lb_metric == 'lbr': # remove zeroes
lbfo = {u : (math.log(v) if args.log_lbr else v) for u, v in lbfo.items() if v > 0}
lbvals = lbfo.values()
if len(lbvals) == 0:
return
lb_smap = plotting.get_normalized_scalar_map(lbvals, 'viridis', hard_min=get_scale_min(args.lb_metric, lbvals) if args.lb_metric=='cons-dist-aa' else None)
lb_min, lb_max = min(lbvals), max(lbvals)
affyfo = None
if args.affy_key in args.metafo and set(args.metafo[args.affy_key].values()) != set([None]):
affyfo = args.metafo[args.affy_key]
if args.lb_metric in affy_metrics:
affyvals = affyfo.values()
affy_smap = plotting.get_normalized_scalar_map([a for a in affyvals if a is not None], 'viridis')
elif args.lb_metric in delta_affy_metrics:
delta_affyvals = set_delta_affinities(etree, affyfo)
delta_affy_increase_smap = plotting.get_normalized_scalar_map([v for v in delta_affyvals if v > 0], 'Reds', remove_top_end=True) if len(delta_affyvals) > 0 else None
delta_affy_decrease_smap = plotting.get_normalized_scalar_map([abs(v) for v in delta_affyvals if v < 0], 'Blues', remove_top_end=True) if len(delta_affyvals) > 0 else None
else:
assert False
for node in etree.traverse():
node.img_style['size'] = 0
rfsize = 0
bgcolor = plotting.getgrey()
if args.lb_metric in affy_metrics:
if node.name not in lbfo: # really shouldn't happen
print ' %s missing lb info for node \'%s\'' % (utils.color('red', 'warning'), node.name)
continue
if affyfo is not None:
rfsize = get_size(lb_min, lb_max, lbfo[node.name])
if node.name in affyfo:
bgcolor = plotting.get_smap_color(affy_smap, affyfo, key=node.name)
else:
rfsize = 5
bgcolor = plotting.get_smap_color(lb_smap, lbfo, key=node.name)
elif args.lb_metric in delta_affy_metrics:
node.img_style['vt_line_color'] = plotting.getgrey() # if they're black, it's too hard to see the large changes in affinity, since they're very dark (at least with current color schemes)
# rfsize = get_size(lb_min, lb_max, lbfo[node.name]) if node.name in lbfo else 1.5
rfsize = 5 if node.name in lbfo else 1.5
bgcolor = plotting.get_smap_color(lb_smap, lbfo, key=node.name)
if affyfo is not None and delta_affy_increase_smap is not None and node.affinity_change is not None:
# tface = ete3.TextFace(('%+.4f' % node.affinity_change) if node.affinity_change != 0 else '0.', fsize=3)
# node.add_face(tface, column=0)
if node.affinity_change > 0: # increase
node.img_style['hz_line_color'] = plotting.get_smap_color(delta_affy_increase_smap, None, val=node.affinity_change)
node.img_style['hz_line_width'] = 1.2
elif node.affinity_change < 0: # decrease
node.img_style['hz_line_color'] = plotting.get_smap_color(delta_affy_decrease_smap, None, val=abs(node.affinity_change))
node.img_style['hz_line_width'] = 1.2
else:
node.img_style['hz_line_color'] = plotting.getgrey()
if args.queries_to_include is not None and node.name in args.queries_to_include:
tface = ete3.TextFace(node.name, fsize=3, fgcolor='red')
node.add_face(tface, column=0)
rface = ete3.RectFace(width=rfsize, height=rfsize, bgcolor=bgcolor, fgcolor=None)
rface.opacity = opacity
node.add_face(rface, column=0)
affy_label = args.affy_key.replace('_', ' ')
if args.lb_metric in affy_metrics:
if affyfo is None:
add_legend(tstyle, args.lb_metric, lbvals, lb_smap, lbfo, 0, n_entries=4)
else:
add_legend(tstyle, args.lb_metric, lbvals, None, lbfo, 0, n_entries=4)
add_legend(tstyle, affy_label, [a for a in affyvals if a is not None], affy_smap, affyfo, 3)
elif args.lb_metric in delta_affy_metrics:
add_legend(tstyle, args.lb_metric, lbvals, lb_smap, lbfo, 0, reverse_log=args.log_lbr)
if affyfo is not None:
add_legend(tstyle, '%s decrease' % affy_label, [abs(v) for v in delta_affyvals if v < 0], delta_affy_decrease_smap, affyfo, 3, add_sign='-', no_opacity=True)
add_legend(tstyle, '%s increase' % affy_label, [v for v in delta_affyvals if v > 0], delta_affy_increase_smap, affyfo, 6, add_sign='+', no_opacity=True)
