def add_branch_text(tree, tree_style, node_dict):
root_node = tree & 'root'
for node in tree.traverse():
if node is root_node:
continue
T1 = TextFace(node.name, ftype="Monaco", fsize=10)
T1.hz_align = 0
node.add_face(T1, 0, 'branch-top')
T2 = TextFace('%s%%' % float_trans(node_dict[node.name].profile * 100), ftype="Monaco", fsize=10)
T2.hz_align = 0
node.add_face(T2, 0, 'branch-bottom')
# print node_dict[node.name].size
node.dist = node_dict[node.name].branch_length
tree_style.scale = 1
def my_layout(node):
circle_color = 'lightgray' if colormap is None or node.name not in colormap else colormap[
node.name]
text_color = 'black'
if isinstance(circle_color, str):
C = CircleFace(radius=max(3, 10 * scipy.sqrt(node.frequency)),
color=circle_color,
label={
'text': str(node.frequency),
'color': text_color
} if node.frequency > 0 else None)
C.rotation = -90
C.hz_align = 1
faces.add_face_to_node(C, node, 0)
else:
P = PieChartFace(
[100 * x / node.frequency for x in circle_color.values()],
2 * 10 * scipy.sqrt(node.frequency),
2 * 10 * scipy.sqrt(node.frequency),
colors=[(color if color != 'None' else 'lightgray')
for color in list(circle_color.keys())],
line_color=None)
T = TextFace(' '.join(
[str(x) for x in list(circle_color.values())]),
tight_text=True)
T.hz_align = 1
T.rotation = -90
faces.add_face_to_node(P, node, 0, position='branch-right')
faces.add_face_to_node(T, node, 1, position='branch-right')
if idlabel:
T = TextFace(node.name, tight_text=True, fsize=6)
T.rotation = -90
T.hz_align = 1
faces.add_face_to_node(
T,
node,
1 if isinstance(circle_color, str) else 2,
position='branch-right')
def _add_header(self,
header_name,
column_add=0):
n = TextFace(f'{header_name}')
n.margin_top = 1
n.margin_right = 1
n.margin_left = 20
n.margin_bottom = 1
n.hz_align = 2
n.vt_align = 2
n.rotation = 270
n.inner_background.color = "white"
n.opacity = 1.
# add header
self.tss.aligned_header.add_face(n, self.column_count-1+column_add)
def traitTree(tree, traits, mapper, outDir, floatSwitch=0):
### Take dict of traits and [R,G,B]-returning function
### Draw a tree with the continuous trait painted on via a colormapping function
def rgb2hex(r, g, b):
hex = "#{:02x}{:02x}{:02x}".format(r, g, b)
return hex
for n in tree.traverse():
if n.is_leaf():
n.set_style(nstyle_L)
# a manual way to set leaf labels
#n.add_face(TextFace(str(n.name)), column=0, position="aligned")
else:
n.set_style(nstyle)
#nd = TextFace(str(n.ND)) # label with node ID
# make the face
if round == 0:
nd = TextFace(str(int(round(
traits[n.ND])))) # label with rounded continuous trait value
else:
nd = CircleFace(
10, color=rgb2hex(*[int(val) for val in mapper(traits[n.ND])]))
#nd = TextFace(str(round(traits[n.ND], floatSwitch))) # label with rounded continuous trait value
#nd.background.color = rgb2hex(*[int(val) for val in mapper(traits[n.ND], gamma=0.8, scaleMax=255)]) # setup for wl2RGB
#nd.background.color = rgb2hex(*[int(val) for val in mapper(traits[n.ND])]) # setup for grayscale
nd.background.color = None
#nd.margin_right = 2
#nd.margin_top = 1
#nd.margin_left = 2
#nd.margin_bottom = 1
nd.border.width = None
nd.hz_align = 2
#nd.inner_border.width = 1
# outline for the circle node
#ol = CircleFace(11, color = "black")
#ol.hz_align = 0
#n.add_face(ol, column=0, position="float-behind") #float-behind
n.add_face(nd, column=0, position="float") #float
# this is necessary for some reason to keep the tree from collapsing
n.add_face(TextFace(" "), column=0, position="branch-bottom")
outFile = outDir + "/cont_trait.pdf"
tree.render(outFile, tree_style=tree_style)
print >> sys.stderr, outFile
def get_tree_style():
ts = TreeStyle()
# ts.mode = 'c'
ts.margin_top = 10
ts.margin_bottom = 10
ts.margin_left = 10
ts.margin_right = 10
ts.show_leaf_name = False
ts.show_branch_length = False
ts.show_branch_support = False
ts.show_scale = False
title = TextFace(" Tax Assignment Tree", fsize=10)
title.hz_align = 2
title.vt_align = 2
ts.title.add_face(TextFace(" "), column=0)
ts.title.add_face(TextFace(" "), column=0)
ts.title.add_face(title, column=0)
return ts
def add_heatmap(self,
taxon2value,
header_name,
continuous_scale=False,
show_text=False):
from metagenlab_libs.colors import get_continuous_scale
self._add_header(header_name)
if continuous_scale:
color_scale = get_continuous_scale(taxon2value.values())
for i, lf in enumerate(self.tree.iter_leaves()):
if not lf.name in taxon2value:
n = TextFace('')
else:
value = taxon2value[lf.name]
if show_text:
n = TextFace('%s' % value)
else:
n = TextFace(' ')
n.margin_top = 2
n.margin_right = 3
n.margin_left = 3
n.margin_bottom = 2
n.hz_align = 1
n.vt_align = 1
n.border.width = 3
n.border.color = "#ffffff"
if continuous_scale:
n.background.color = rgb2hex(color_scale[0].to_rgba(float(value)))
n.opacity = 1.
i+=1
if self.rotate:
n.rotation = 270
lf.add_face(n, self.column_count, position="aligned")
self.column_count += 1
def my_layout(node):
circle_color = 'lightgray' if colormap is None or node.name not in colormap else colormap[
node.name]
text_color = 'black'
if isinstance(circle_color, str):
if isolabel and hasattr(node, 'isotype'):
nl = ''.join(
sorted(set([ISO_SHORT[iss] for iss in node.isotype]),
key=lambda x: ISO_TYPE_charORDER[x]))
else:
nl = str(node.frequency)
C = CircleFace(radius=max(3, 10 * scipy.sqrt(node.frequency)),
color=circle_color,
label={
'text': nl,
'color': text_color
} if node.frequency > 0 else None)
C.rotation = -90
C.hz_align = 1
faces.add_face_to_node(C, node, 0)
else:
P = PieChartFace(
[100 * x / node.frequency for x in circle_color.values()],
2 * 10 * scipy.sqrt(node.frequency),
2 * 10 * scipy.sqrt(node.frequency),
colors=[(color if color != 'None' else 'lightgray')
for color in list(circle_color.keys())],
line_color=None)
T = TextFace(' '.join(
[str(x) for x in list(circle_color.values())]),
tight_text=True)
T.hz_align = 1
T.rotation = -90
faces.add_face_to_node(P, node, 0, position='branch-right')
faces.add_face_to_node(T, node, 1, position='branch-right')
if idlabel:
T = TextFace(node.name, tight_text=True, fsize=6)
T.rotation = -90
T.hz_align = 1
faces.add_face_to_node(
T,
node,
1 if isinstance(circle_color, str) else 2,
position='branch-right')
elif isolabel and hasattr(node, 'isotype') and False:
iso_name = ''.join(
sorted(set([ISO_SHORT[iss] for iss in node.isotype]),
key=lambda x: ISO_TYPE_charORDER[x]))
#T = TextFace(iso_name, tight_text=True, fsize=6)
#T.rotation = -90
#T.hz_align = 1
#faces.add_face_to_node(T, node, 1 if isinstance(circle_color, str) else 2, position='branch-right')
C = CircleFace(radius=max(3, 10 * scipy.sqrt(node.frequency)),
color=circle_color,
label={
'text': iso_name,
'color': text_color
} if node.frequency > 0 else None)
C.rotation = -90
C.hz_align = 1
faces.add_face_to_node(C, node, 0)
def heatmap_view(tree, orthologous_groups, save_dir):
"""Generates a heatmap of regulation states in all species."""
light_tree = copy.deepcopy(tree) # Tree copy for the light heatmap
# Heat map settings
rect_face_fgcolor = 'black'
locus_tag_len = max(
len(gene.locus_tag) + 5 for ortho_grp in orthologous_groups
for gene in ortho_grp.genes)
rect_face_width = locus_tag_len * 8
light_rect_face_width = 20
rect_face_height = 20
rotation = 90
# Sort orthologous groups by the number of regulated genes in each group
orthologous_groups = filter_and_sort_orthologous_grps(orthologous_groups)
# For each species and its gene in each orthologous group, draw a rectangle
for node, light_node in zip(tree.get_leaves(), light_tree.get_leaves()):
for i, orthologous_grp in enumerate(orthologous_groups, start=1):
#get all orthologs in group
matching_genes = [g for g in orthologous_grp.genes \
if g.genome.strain_name == node.name]
#if there is ortholog
if len(matching_genes) > 0:
# Get the first ortholog from the genome in the group
#this is the one with higher probability of regulation.
#so this probability will be displayed for the group
gene = matching_genes[0]
p_regulation = gene.operon.regulation_probability
p_notregulation = 1.0 - p_regulation
p_absence = 0
# No ortholog from this genome
else:
gene = None
p_regulation = 0
p_notregulation = 0
p_absence = 1
# Color of the rectangle is based on probabilities
rect_face_bgcolor = rgb2hex(p_notregulation, p_regulation,
p_absence)
rect_face_text = ('%s [%d]' %
(gene.locus_tag, gene.operon.operon_id)
if gene else '')
rect_face_label = {
'text': rect_face_text,
'font': 'Courier',
'fontsize': 8,
'color': 'black'
}
# Create the rectangle
rect_face = RectFace(rect_face_width,
rect_face_height,
rect_face_fgcolor,
rect_face_bgcolor,
label=rect_face_label)
light_rect_face = RectFace(light_rect_face_width,
rect_face_height,
rect_face_fgcolor,
rect_face_bgcolor,
label='')
rect_face.rotation = -rotation
light_rect_face.rotation = -rotation
# Add the rectangle to the corresponding column
node.add_face(rect_face, column=i, position='aligned')
light_node.add_face(light_rect_face, column=i, position='aligned')
ts = TreeStyle()
# Add orthologous group descriptions
descriptions = ['-'.join([grp.description, \
str([item['ID'] for item in grp.COGs]) if len(grp.COGs)>0 else '', \
str([item['ID'] for item in grp.NOGs]) if len(grp.NOGs)>0 else '', \
str([item['ID'] for item in grp.PFAMs])] if len(grp.PFAMs)>0 else '')\
for grp in orthologous_groups]
max_description_len = max(map(len, descriptions))
descriptions = [
'[%d]' % i + description + ' ' *
(max_description_len - len(description))
for i, description in enumerate(descriptions, start=1)
]
for i, description in enumerate(descriptions, start=1):
text_face = TextFace(description, ftype='Courier')
text_face.hz_align = 1
text_face.vt_align = 1
text_face.rotation = -rotation
ts.aligned_header.add_face(text_face, column=i)
# Rotate the generated heatmap.
ts.margin_left = 10
ts.margin_top = 20
ts.rotation = rotation
ts.show_scale = False
# For some reason, it can't render to PDF in color
tree.render(os.path.join(save_dir, 'heatmap.svg'), tree_style=ts)
light_tree.render(os.path.join(save_dir, 'heatmap_light.svg'),
tree_style=ts)
def generateFigure(PF, sample, rank, input_file, output_base_name, file_type, plot_l1, scaling, output_dpi):
# Make the ETE3 tree
try:
tree = ncbi.get_topology(PF.get_all_tax_ids(sample), rank_limit=rank)
except:
logging.getLogger('Tampa').critical("Input format not compatible.")
exit(1)
ts = TreeStyle()
ts.layout_fn = PF.layout
ts.mode = "c"
ts.show_leaf_name = False
ts.show_branch_length = False
ts.show_branch_support = False
ts.min_leaf_separation = 10
ts.arc_span = 360
#ts.legend.add_face(CircleFace(100, "#1b9e77", label="Predicted"), column=0)
#ts.legend.add_face(CircleFace(100, '#d95f02', label="True"), column=1)
# add white space to move the legend closer
ts.legend.add_face(CircleFace(65, "#FFFFFF"), column=2)
ts.legend.add_face(CircleFace(65, "#FFFFFF"), column=1)
ts.legend.add_face(CircleFace(65, "#FFFFFF"), column=0)
ts.legend.add_face(CircleFace(65, "#FFFFFF"), column=2)
ts.legend.add_face(CircleFace(65, "#FFFFFF"), column=1)
ts.legend.add_face(CircleFace(65, "#FFFFFF"), column=0)
# add the legend
legend_fs = 50
C1 = CircleFace(100, "#1b9e77")
C1.hz_align = True
ts.legend.add_face(C1, column=0)
T1 = TextFace("Predicted", fsize=legend_fs)
T1.hz_align = True
ts.legend.add_face(T1, column=0)
if len(PF.ground_truth_dict) > 0:
C2 = CircleFace(100, "#d95f02")
C2.hz_align = True
ts.legend.add_face(C2, column=1)
T2 = TextFace("True", fsize=legend_fs)
T2.hz_align = True
ts.legend.add_face(T2, column=1)
T3 = TextFace(f"Tool: {os.path.basename(input_file).split('.')[0]}", fsize=legend_fs)
T3.hz_align = True
ts.legend.add_face(T3, column=0)
ts.allow_face_overlap = False # this lets me mess a bit with font size and face size without the interaction of the two
ts.min_leaf_separation = 10
tree_output_file = f"{output_base_name}_tree_{rank}_{sample}.{file_type}"
tree.render(tree_output_file, h=5.2, w=5, tree_style=ts, units="in", dpi=output_dpi)
if plot_l1:
# if you asked for L1 too, then plot that
true_abundance_at_rank = []
predicted_abundance_at_rank = []
for node in tree.get_leaves():
if node.rank == rank:
tax_id = str(node.taxid)
if tax_id in PF.ground_truth_tax_id_to_percentage:
true_abundance_at_rank.append(PF.ground_truth_tax_id_to_percentage[str(node.taxid)] / 100.)
else:
true_abundance_at_rank.append(0)
if tax_id in PF.profile_tax_id_to_percentage:
predicted_abundance_at_rank.append(PF.profile_tax_id_to_percentage[str(node.taxid)] / 100.)
else:
predicted_abundance_at_rank.append(0)
data = np.zeros((len(true_abundance_at_rank), 2))
data[:, 0] = np.array(true_abundance_at_rank)
data[:, 1] = np.array(predicted_abundance_at_rank)
df = pd.DataFrame(data, columns=['True', 'Predicted'])
# g = seaborn.FacetGrid(df, height=6)
ax = seaborn.scatterplot(x='True', y='Predicted', data=df, color='b', s=55)
eps = 1
ax.set_aspect('equal')
max_val = np.max(data) + eps
ax.set_xlim(-.5, max_val)
ax.set_ylim(-.5, max_val)
ax.set_xbound(-.5, max_val)
ax.set_ybound(-.5, max_val)
#plt.figure(figsize=(6,6))
plt.plot(np.linspace(0, max_val, 100), np.linspace(0, max_val, 100), color='k')
for (x, y) in zip(true_abundance_at_rank, predicted_abundance_at_rank):
if x > y:
ax.vlines(x, y, x, colors='r')
if y > x:
ax.vlines(x, x, y, colors='r')
plt.title(f"Tool: {os.path.basename(input_file).split('.')[0]}")
plt.tight_layout()
l1_out_file = f"{output_base_name}_L1_{rank}.{file_type}"
plt.savefig(l1_out_file, dpi=output_dpi)
def plot_phylum_counts(NOG_id,
rank='phylum',
colapse_low_species_counts=4,
remove_unlassified=True):
'''
1. get phylum tree
2. foreach species => get phylum
3. build phylum2count dictionnary
3. plot barchart
# merge eukaryotes into 5 main clades
# merge virus as a single clade
ATTENTION: no-rank groups and no-rank species...
'''
import MySQLdb
import os
from chlamdb.biosqldb import manipulate_biosqldb
from ete3 import NCBITaxa, Tree, TextFace, TreeStyle, StackedBarFace
ncbi = NCBITaxa()
sqlpsw = os.environ['SQLPSW']
conn = MySQLdb.connect(
host="localhost", # your host, usually localhost
user="******", # your username
passwd=sqlpsw, # your password
db="eggnog") # name of the data base
cursor = conn.cursor()
sql = 'select * from eggnog.leaf2n_genomes_%s' % rank
cursor.execute(sql, )
leaf_taxon2n_species = manipulate_biosqldb.to_dict(cursor.fetchall())
leaf_taxon2n_species_with_domain = get_NOG_taxonomy(NOG_id, rank)
sql = 'select phylogeny from eggnog.phylogeny where rank="%s"' % (rank)
cursor.execute(sql, )
tree = Tree(cursor.fetchall()[0][0], format=1)
sql = 'select * from eggnog.taxid2label_%s' % rank
cursor.execute(sql, )
taxon_id2scientific_name_and_rank = manipulate_biosqldb.to_dict(
cursor.fetchall())
taxon_id2scientific_name_and_rank = {
str(k): v
for k, v in taxon_id2scientific_name_and_rank.items()
}
tss = TreeStyle()
tss.draw_guiding_lines = True
tss.guiding_lines_color = "blue"
keep = []
for lf in tree.iter_leaves():
# n genomes
if remove_unlassified:
label = taxon_id2scientific_name_and_rank[str(lf.name)][0]
if 'unclassified' in label:
continue
n_genomes = int(leaf_taxon2n_species[lf.name])
if n_genomes > colapse_low_species_counts:
keep.append(lf.name)
print('number of leaaves:', len(keep))
tree.prune(keep)
header_list = ['Rank', 'N genomes', 'N with %s' % NOG_id, 'Percentage']
for col, header in enumerate(header_list):
n = TextFace('%s' % (header))
n.margin_top = 0
n.margin_right = 1
n.margin_left = 20
n.margin_bottom = 1
n.rotation = 270
n.hz_align = 2
n.vt_align = 2
n.inner_background.color = "white"
n.opacity = 1.
tss.aligned_header.add_face(n, col)
for lf in tree.iter_leaves():
# n genomes
n_genomes = int(leaf_taxon2n_species[lf.name])
if n_genomes <= colapse_low_species_counts:
continue
n = TextFace(' %s ' % str(leaf_taxon2n_species[lf.name]))
n.margin_top = 1
n.margin_right = 1
n.margin_left = 0
n.margin_bottom = 1
n.fsize = 7
n.inner_background.color = "white"
n.opacity = 1.
lf.add_face(n, 2, position="aligned")
# n genomes with domain
try:
m = TextFace(' %s ' %
str(leaf_taxon2n_species_with_domain[lf.name]))
except:
m = TextFace(' 0 ')
m.margin_top = 1
m.margin_right = 1
m.margin_left = 0
m.margin_bottom = 1
m.fsize = 7
m.inner_background.color = "white"
m.opacity = 1.
lf.add_face(m, 3, position="aligned")
# rank
ranks = ncbi.get_rank([lf.name])
try:
r = ranks[max(ranks.keys())]
except:
r = '-'
n = TextFace(' %s ' % r, fsize=14, fgcolor='red')
n.margin_top = 1
n.margin_right = 1
n.margin_left = 0
n.margin_bottom = 1
n.fsize = 7
n.inner_background.color = "white"
n.opacity = 1.
lf.add_face(n, 1, position="aligned")
# percent with target domain
try:
percentage = (float(leaf_taxon2n_species_with_domain[lf.name]) /
float(leaf_taxon2n_species[lf.name])) * 100
except:
percentage = 0
m = TextFace(' %s ' % str(round(percentage, 2)))
m.fsize = 1
m.margin_top = 1
m.margin_right = 1
m.margin_left = 0
m.margin_bottom = 1
m.fsize = 7
m.inner_background.color = "white"
m.opacity = 1.
lf.add_face(m, 4, position="aligned")
b = StackedBarFace([percentage, 100 - percentage],
width=100,
height=10,
colors=["#7fc97f", "white"])
b.rotation = 0
b.inner_border.color = "grey"
b.inner_border.width = 0
b.margin_right = 15
b.margin_left = 0
lf.add_face(b, 5, position="aligned")
n = TextFace('%s' % taxon_id2scientific_name_and_rank[str(lf.name)][0],
fgcolor="black",
fsize=9) # , fstyle = 'italic'
lf.name = " %s (%s)" % (taxon_id2scientific_name_and_rank[str(
lf.name)][0], str(lf.name))
n.margin_right = 10
lf.add_face(n, 0)
tss.show_leaf_name = False
for node in tree.traverse("postorder"):
try:
r = taxon_id2scientific_name_and_rank[str(node.name)][1]
except:
pass
try:
if r in ['phylum', 'superkingdom', 'class', 'subphylum'
] or taxon_id2scientific_name_and_rank[str(
node.name)][0] in ['FCB group']:
hola = TextFace(
"%s" %
(taxon_id2scientific_name_and_rank[str(node.name)][0]))
node.add_face(hola, column=0, position="branch-top")
except:
pass
return tree, tss
def make_cluster_tree(tree_file: str,
matrix: str,
out_file: str,
outgroup: Optional[List[str]] = None) -> None:
"""Draw a tree with cluster absence/presence information from an existing
tree file and absence/presence matrix, and save it as an image under the
supplied file name.
Arguments:
tree_file: the name of the file containing the tree to annotate
matrix: a comma- or tab-separated absence/presence matrix
out_file: the name under which to save the resulting image
outgroup: the organism(s) to use as an outgroup, if any
"""
# ClusterTree needs tab-separated, but that can't be exported cleanly
matrix = matrix.replace(",", "\t")
# tree with clustering analysis
tree = ClusterTree(tree_file, text_array=matrix)
# rerooting the tree
if outgroup:
ancestor = tree.get_common_ancestor(outgroup)
tree.set_outgroup(ancestor)
tree.ladderize(direction=1)
# set drawing line width to 2
my_node_style = NodeStyle()
my_node_style["vt_line_width"] = 2
my_node_style["hz_line_width"] = 2
my_node_style["size"] = 5
# layout function
def sel_mylayout(node):
node.set_style(my_node_style)
if node.is_leaf():
# add names in larger font + italics
species_name = AttrFace("name", fsize=12, fstyle="italic")
add_face_to_node(species_name,
node,
column=0,
position="branch-right")
# add absence/presence matrix
for i, value in enumerate(getattr(node, "profile", [])):
if value > 0:
color = "#FF0000"
else:
color = "#EEEEEE"
my_face = CircleFace(8, color, style="circle")
my_face.margin_right = 3
my_face.margin_bottom = 3
add_face_to_node(my_face, node, position="aligned", column=i)
# Use my layout to visualize the tree
my_tree_style = TreeStyle()
# Add header
for j, name in enumerate(tree.arraytable.colNames):
name_face = TextFace(name, fsize=11)
name_face.rotation = -90
name_face.hz_align = 1
name_face.vt_align = 1
name_face.margin_bottom = 10
my_tree_style.aligned_header.add_face(name_face, column=j)
my_tree_style.scale_length = 0.1
# myTreeStyle.show_branch_support = True
# don't auto-show leaf names, since we dealt with that above
my_tree_style.show_leaf_name = False
# set layout function for my_tree_style
my_tree_style.layout_fn = sel_mylayout
#tree.render(out_file, w=183, units="mm", dpi=600, tree_style=my_tree_style)
tree.render(out_file, dpi=600, tree_style=my_tree_style)
def plot_tree_barplot(tree_file, taxon2mlst, header_list):
'''
display one or more barplot
:param tree_file:
:param taxon2value_list:
:param exclude_outgroup:
:param bw_scale:
:param barplot2percentage: list of bool to indicates if the number are percentages and the range should be set to 0-100
:return:
'''
import matplotlib.cm as cm
from matplotlib.colors import rgb2hex
import matplotlib as mpl
mlst_list = list(set(taxon2mlst.values()))
mlst2color = dict(zip(mlst_list, get_spaced_colors(len(mlst_list))))
mlst2color['-'] = 'white'
if isinstance(tree_file, Tree):
t1 = tree_file
else:
t1 = Tree(tree_file)
# Calculate the midpoint node
R = t1.get_midpoint_outgroup()
# and set it as tree outgroup
t1.set_outgroup(R)
tss = TreeStyle()
value = 1
tss.draw_guiding_lines = True
tss.guiding_lines_color = "gray"
tss.show_leaf_name = False
cmap = cm.YlGnBu #YlOrRd#OrRd
scale_list = []
max_value_list = []
for i, lf in enumerate(t1.iter_leaves()):
#if taxon2description[lf.name] == 'Pirellula staleyi DSM 6068':
# lf.name = 'Pirellula staleyi DSM 6068'
# continue
if i == 0:
# header
col_add = 0
#lf.add_face(n, column, position="aligned")
n = TextFace('MLST')
n.margin_top = 1
n.margin_right = 2
n.margin_left = 2
n.margin_bottom = 1
n.rotation = 90
n.inner_background.color = "white"
n.opacity = 1.
n.hz_align = 2
n.vt_align = 2
tss.aligned_header.add_face(n, col_add + 1)
try:
#if lf.name in leaf2mlst or int(lf.name) in leaf2mlst:
n = TextFace(' %s ' % taxon2mlst[int(lf.name)])
n.inner_background.color = 'white'
m = TextFace(' ')
m.inner_background.color = mlst2color[taxon2mlst[int(lf.name)]]
except:
n = TextFace(' na ')
n.inner_background.color = "grey"
m = TextFace(' ')
m.inner_background.color = "white"
n.opacity = 1.
n.margin_top = 2
n.margin_right = 2
n.margin_left = 0
n.margin_bottom = 2
m.margin_top = 2
m.margin_right = 0
m.margin_left = 2
m.margin_bottom = 2
lf.add_face(m, 0, position="aligned")
lf.add_face(n, 1, position="aligned")
n = TextFace(lf.name, fgcolor="black", fsize=12, fstyle='italic')
lf.add_face(n, 0)
for n in t1.traverse():
nstyle = NodeStyle()
if n.support < 1:
nstyle["fgcolor"] = "black"
nstyle["size"] = 6
n.set_style(nstyle)
else:
nstyle["fgcolor"] = "red"
nstyle["size"] = 0
n.set_style(nstyle)
return t1, tss
def plot_ete_tree(tree_file,
ordered_queries,
leaf_id2protein_id2identity,
leaf_id2mlst,
leaf_id2spa,
leaf_id2meca,
show_identity_values=True,
leaf_id2description=False):
mlst_list = list(set(leaf_id2mlst.values()))
mlst2color = dict(zip(mlst_list, get_spaced_colors(len(mlst_list))))
mlst2color['-'] = 'white'
t1 = Tree(tree_file)
tss = TreeStyle()
R = t1.get_midpoint_outgroup()
t1.set_outgroup(R)
t1.ladderize()
head = True
column_add = 4
for lf in t1.iter_leaves():
lf.branch_vertical_margin = 0
# add MLST
if head:
n = TextFace(' MLST ')
n.margin_top = 2
n.margin_right = 2
n.margin_left = 2
n.margin_bottom = 2
n.rotation = 270
n.vt_align = 2
n.hz_align = 2
n.inner_background.color = "white"
n.opacity = 1.
tss.aligned_header.add_face(n, 1)
if lf.name in leaf2mlst:
n = TextFace(' %s ' % leaf_id2mlst[lf.name])
n.inner_background.color = 'white'
m = TextFace(' ')
m.inner_background.color = mlst2color[leaf_id2mlst[lf.name]]
else:
n = TextFace(' na ')
n.inner_background.color = "grey"
m = TextFace(' ')
m.inner_background.color = "white"
n.opacity = 1.
n.margin_top = 2
n.margin_right = 2
n.margin_left = 0
n.margin_bottom = 2
m.margin_top = 2
m.margin_right = 0
m.margin_left = 20
m.margin_bottom = 2
lf.add_face(m, 0, position="aligned")
lf.add_face(n, 1, position="aligned")
# add spa typing
if head:
n = TextFace(' spa ')
n.margin_top = 2
n.margin_right = 2
n.margin_left = 2
n.margin_bottom = 2
n.rotation = 270
n.vt_align = 2
n.hz_align = 2
n.inner_background.color = "white"
n.opacity = 1.
tss.aligned_header.add_face(n, column_add-2)
if lf.name in leaf_id2spa:
n = TextFace(' %s ' % leaf_id2spa[lf.name])
n.inner_background.color = "white"
else:
n = TextFace(' na ')
n.inner_background.color = "grey"
n.opacity = 1.
n.margin_top = 2
n.margin_right = 2
n.margin_left = 2
n.margin_bottom = 2
lf.add_face(n, column_add-2, position="aligned")
# add mecA typing
if head:
n = TextFace(' mecA ')
n.margin_top = 2
n.margin_right = 2
n.margin_left = 2
n.margin_bottom = 2
n.rotation = 270
n.vt_align = 2
n.hz_align = 2
n.inner_background.color = "white"
n.opacity = 1.
tss.aligned_header.add_face(n, column_add-1)
if lf.name in leaf_id2meca:
n = TextFace(' %s ' % leaf_id2meca[lf.name])
if leaf_id2meca[lf.name] == 'Perfect':
n.inner_background.color = "red"
elif leaf_id2meca[lf.name] == 'Strict':
n.inner_background.color = "orange"
else:
n.inner_background.color = "white"
else:
n = TextFace(' na ')
n.inner_background.color = "grey"
n.opacity = 1.
n.margin_top = 2
n.margin_right = 2
n.margin_left = 2
n.margin_bottom = 2
lf.add_face(n, column_add-1, position="aligned")
# loop to add virulence gene hits
for column, protein_id in enumerate(ordered_queries):
# draw labels at the top of each column
if head:
if show_identity_values:
n = TextFace(' %s ' % str(protein_id))
n.margin_top = 2
n.margin_right = 2
n.margin_left = 2
n.margin_bottom = 2
n.rotation = 270
n.vt_align = 2
n.hz_align = 2
n.inner_background.color = "white"
n.opacity = 1.
tss.aligned_header.add_face(n, column+column_add)
else:
n = TextFace(' %s ' % str(protein_id), fsize=6)
n.margin_top = 0
n.margin_right = 0
n.margin_left = 0
n.margin_bottom = 0
n.rotation = 270
n.vt_align = 2
n.hz_align = 2
n.inner_background.color = "white"
n.opacity = 1.
# lf.add_face(n, col, position="aligned")
tss.aligned_header.add_face(n, column+column_add)
# draw column content
if lf.name not in leaf_id2protein_id2identity:
n = TextFace(' %s ' % str(' na '))
n.opacity = 1.
n.margin_top = 2
n.margin_right = 2
n.margin_left = 2
n.margin_bottom = 2
n.inner_background.color = "grey"
lf.add_face(n, column+column_add, position="aligned")
else:
if protein_id in leaf_id2protein_id2identity[lf.name]:
identity_value = float(leaf_id2protein_id2identity[lf.name][protein_id])
color = rgb2hex(m_blue.to_rgba(identity_value))
if show_identity_values:
# report identity values in coloured boxes
# adapt box size depending the digit width
if str(identity_value) == '100.00' or str(identity_value) == '100.0':
identity_value = '100'
n = TextFace(" %s " % identity_value)
else:
n = TextFace("%.2f" % round(float(identity_value), 2))
# color text to white for dark cells
if float(identity_value) > 95:
n.fgcolor = "white"
n.opacity = 1.
n.margin_top = 2
n.margin_right = 2
n.margin_left = 2
n.margin_bottom = 2
n.inner_background.color = color
lf.add_face(n, column+column_add, position="aligned")
else:
# draw coloured boxes without text
n = TextFace(' ')
n.margin_top = 0
n.margin_right = 0
n.margin_left = 0
n.margin_bottom = 0
# n.color = color
n.inner_background.color = color
lf.add_face(n, column+column_add, position="aligned")
else:
n = TextFace(' %s ' % str(' - '))
n.opacity = 1.
n.margin_top = 2
n.margin_right = 2
n.margin_left = 2
n.margin_bottom = 2
n.inner_background.color = "white"
lf.add_face(n, column+column_add, position="aligned")
# end of first leaf: turn off header
head = False
# add boostrap supports
for n in t1.traverse():
nstyle = NodeStyle()
if n.support < 0.9:
nstyle["fgcolor"] = "blue"
nstyle["size"] = 6
n.set_style(nstyle)
else:
nstyle["fgcolor"] = "red"
nstyle["size"] = 0
n.set_style(nstyle)
return t1, tss
def custom_layout(self,node):
if node.is_leaf():
aligned_name_face = TextFace(node.name, fgcolor='olive', fsize=12)
aligned_name_face.margin_top = 5
aligned_name_face.margin_right = 5
aligned_name_face.margin_left = 5
aligned_name_face.margin_bottom = 5
aligned_name_face.hz_align = 0 #0 = left, 1 = center, 2 = right
add_face_to_node(aligned_name_face, node, column=2, position='aligned')
#name_face = TextFace(node.name, fgcolor='#333333', fsize=11)
#name_face.margin_top = 3
#name_face.margin_right = 3
#name_face.margin_left = 3
#name_face.margin_bottom = 3
#add_face_to_node(name_face, node, column=2, position='branch-right')
node.img_style['size'] = 0
#---------------------------------------------
#displaying extra categorical and numeric data
if (node.name in self._tip2info):
column_no = 3
for headerIndex, dataheader in enumerate(self._tip2headers):
extra_data = self._tip2info[node.name][headerIndex]
if isinstance( extra_data, ( int, float ) ):
extra_face = BarChartFace([extra_data], width=100,height=25,colors=[self._tip2color[node.name][headerIndex]],labels=[dataheader],min_value=0.0,max_value=self._tip_max)
else:
extra_face = TextFace(extra_data, fsize=11, fgcolor='black')
extra_face.background.color = self._tip2color[node.name][headerIndex]
extra_face.margin_left = 5
extra_face.margin_top = 5
extra_face.margin_right = 5
extra_face.margin_bottom = 5
add_face_to_node(extra_face, node, column=column_no, position='aligned')
#add_face_to_node(extra_face, node, column=column_no, aligned = True, position='branch-right')
column_no += 1
else:
#print "No data available"
column_no = 3
for headerIndex, dataheader in enumerate(self._tip2headers):
extra_face = TextFace("No data available", fsize=10, fgcolor='black')
extra_face.margin_left = 5
extra_face.margin_top = 5
extra_face.margin_right = 5
extra_face.margin_bottom = 5
add_face_to_node(extra_face, node, column=column_no, position='aligned')
column_no += 1
image_col_no = column_no
#----------------------------------------------
if (node.name in self._img_chk_list):
if self._img_data_dic[node.name] is not None:
img_face = ImgFace(self._img_data_dic[node.name], is_url=True)
#img_face = ImgFace(self._tip2info[node.name][0], is_url=True)
#img_path = os.path.join("file:///home/tayeen/TayeenFolders/TreeViewer/WebTreeApp/newplugin_test/data/", "328653.jpg")
#img_face = ImgFace(img_path, is_url=True)
img_face.margin_top = 10
img_face.margin_right = 10
img_face.margin_left = 10
img_face.margin_bottom = 10
#add_face_to_node(img_face, node, column=3, position='branch-right')
#add_face_to_node(img_face, node, column=3, aligned= True, position='branch-right')
else:
img_path = os.path.join("file://"+image_path, "ina.jpg")
img_face = ImgFace(img_path, is_url=True)
#add_face_to_node(img_face, node, column=5, position='branch-right')
add_face_to_node(img_face, node, column=image_col_no, position='aligned')
else: #node is not a leaf
node.img_style['size'] = 4
node.img_style['shape'] = 'square'
if node.name and self._custom_options["draw_internal"]:
name_face = TextFace(node.name, fgcolor='grey', fsize=10)
name_face.margin_top = 4
name_face.margin_right = 4
name_face.margin_left = 4
name_face.margin_bottom = 4
add_face_to_node(name_face, node, column=0, position='branch-top')
if node.name in self._node2label:
label_face = TextFace(self._node2label[node.name], fgcolor='DarkGreen', fsize=10)
label_face.margin_top = 4
label_face.margin_right = 4
label_face.margin_left = 4
label_face.margin_bottom = 4
add_face_to_node(label_face, node, column=0, position="branch-top")
if node.support and self._custom_options["draw_support"]:
support_face = TextFace(node.support, fgcolor='indianred', fsize=10)
support_face.margin_top = 4
support_face.margin_right = 4
support_face.margin_left = 4
support_face.margin_bottom = 4
add_face_to_node(support_face, node, column=0, position='branch-bottom')
if hasattr(node, "hide") and int(node.hide) == 1:
node.img_style["draw_descendants"]= False
collapsed_face = faces.TextFace(" %s collapsed leaves." %len(node), \
fsize=10, fgcolor="#444", ftype="Arial")
faces.add_face_to_node(collapsed_face, node, 0)
else:
node.img_style["draw_descendants"] = True
# Parse node features features and conver them into styles. This must be done like this, since current ete version
#does not allow modifying style outside the layout function.
if hasattr(node, "bsize"):
node.img_style["size"]= int(node.bsize)
if hasattr(node, "shape"):
node.img_style["shape"]= node.shape
if hasattr(node, "bgcolor"):
node.img_style["bgcolor"]= node.bgcolor
if hasattr(node, "fgcolor"):
node.img_style["fgcolor"]= node.fgcolor
#parse all nodes features
if hasattr(node, "bh_bgcolor"):
node.img_style["bgcolor"]= node.bh_bgcolor
if hasattr(node, "bh_size"):
node.img_style["size"]= node.bh_size
if hasattr(node, "lh_color"):
node.img_style['hz_line_color'] = node.lh_color
node.img_style["vt_line_color"] = node.lh_color
if hasattr(node, "lh_width"):
node.img_style['hz_line_width'] = node.lh_width
node.img_style['vt_line_width'] = node.lh_width
if hasattr(node, "lh_width") and hasattr(node, "lh_color"):
for n in node.iter_descendants():
n.img_style['hz_line_color'] = node.lh_color
n.img_style["vt_line_color"] = node.lh_color
n.img_style['hz_line_width'] = node.lh_width
n.img_style['vt_line_width'] = node.lh_width
def plot_tree_stacked_barplot(
tree_file,
taxon2value_list_barplot=False,
header_list=False, # header stackedbarplots
taxon2set2value_heatmap=False,
taxon2label=False,
header_list2=False, # header counts columns
biodb=False,
column_scale=True,
general_max=False,
header_list3=False,
set2taxon2value_list_simple_barplot=False,
set2taxon2value_list_simple_barplot_counts=True,
rotate=False,
taxon2description=False):
'''
taxon2value_list_barplot list of lists:
[[bar1_part1, bar1_part2,...],[bar2_part1, bar2_part2]]
valeures de chaque liste transformes en pourcentages
:param tree_file:
:param taxon2value_list:
:param biodb:
:param exclude_outgroup:
:param bw_scale:
:return:
'''
if biodb:
from chlamdb.biosqldb import manipulate_biosqldb
server, db = manipulate_biosqldb.load_db(biodb)
taxon2description = manipulate_biosqldb.taxon_id2genome_description(
server, biodb, filter_names=True)
t1 = Tree(tree_file)
# Calculate the midpoint node
R = t1.get_midpoint_outgroup()
# and set it as tree outgroup
t1.set_outgroup(R)
colors2 = [
"red", "#FFFF00", "#58FA58", "#819FF7", "#F781F3", "#2E2E2E",
"#F7F8E0", 'black'
]
colors = [
"#7fc97f", "#386cb0", "#fdc086", "#ffffb3", "#fdb462", "#f0027f",
"#F7F8E0", 'black'
] # fdc086ff 386cb0ff f0027fff
tss = TreeStyle()
tss.draw_guiding_lines = True
tss.guiding_lines_color = "gray"
tss.show_leaf_name = False
if column_scale and header_list2:
import matplotlib.cm as cm
from matplotlib.colors import rgb2hex
import matplotlib as mpl
column2scale = {}
col_n = 0
for column in header_list2:
values = taxon2set2value_heatmap[column].values()
#print values
if min(values) == max(values):
min_val = 0
max_val = 1.5 * max(values)
else:
min_val = min(values)
max_val = max(values)
#print 'min-max', min_val, max_val
norm = mpl.colors.Normalize(vmin=min_val, vmax=max_val) # *1.1
if col_n < 4:
cmap = cm.OrRd #
else:
cmap = cm.YlGnBu #PuBu#OrRd
m = cm.ScalarMappable(norm=norm, cmap=cmap)
column2scale[column] = [m, float(max_val)] # *0.7
col_n += 1
for i, lf in enumerate(t1.iter_leaves()):
#if taxon2description[lf.name] == 'Pirellula staleyi DSM 6068':
# lf.name = 'Pirellula staleyi DSM 6068'
# continue
if i == 0:
if taxon2label:
n = TextFace(' ')
n.margin_top = 1
n.margin_right = 1
n.margin_left = 20
n.margin_bottom = 1
n.hz_align = 2
n.vt_align = 2
n.rotation = 270
n.inner_background.color = "white"
n.opacity = 1.
tss.aligned_header.add_face(n, 0)
col_add = 1
else:
col_add = 1
if header_list:
for col, header in enumerate(header_list):
n = TextFace('%s' % (header))
n.margin_top = 0
n.margin_right = 1
n.margin_left = 20
n.margin_bottom = 1
n.rotation = 270
n.hz_align = 2
n.vt_align = 2
n.inner_background.color = "white"
n.opacity = 1.
tss.aligned_header.add_face(n, col + col_add)
col_add += col + 1
if header_list3:
#print 'header_list 3!'
col_tmp = 0
for header in header_list3:
n = TextFace('%s' % (header))
n.margin_top = 1
n.margin_right = 1
n.margin_left = 20
n.margin_bottom = 1
n.rotation = 270
n.hz_align = 2
n.vt_align = 2
n.inner_background.color = "white"
n.opacity = 1.
if set2taxon2value_list_simple_barplot_counts:
if col_tmp == 0:
col_tmp += 1
tss.aligned_header.add_face(n, col_tmp + 1 + col_add)
n = TextFace(' ')
tss.aligned_header.add_face(n, col_tmp + col_add)
col_tmp += 2
else:
tss.aligned_header.add_face(n, col_tmp + col_add)
col_tmp += 1
if set2taxon2value_list_simple_barplot_counts:
col_add += col_tmp
else:
col_add += col_tmp
if header_list2:
for col, header in enumerate(header_list2):
n = TextFace('%s' % (header))
n.margin_top = 1
n.margin_right = 1
n.margin_left = 20
n.margin_bottom = 1
n.rotation = 270
n.hz_align = 2
n.vt_align = 2
n.inner_background.color = "white"
n.opacity = 1.
tss.aligned_header.add_face(n, col + col_add)
col_add += col + 1
if taxon2label:
try:
n = TextFace('%s' % taxon2label[lf.name])
except:
try:
n = TextFace('%s' % taxon2label[int(lf.name)])
except:
n = TextFace('-')
n.margin_top = 1
n.margin_right = 1
n.margin_left = 20
n.margin_bottom = 1
n.inner_background.color = "white"
n.opacity = 1.
if rotate:
n.rotation = 270
lf.add_face(n, 1, position="aligned")
col_add = 2
else:
col_add = 2
if taxon2value_list_barplot:
try:
val_list_of_lists = taxon2value_list_barplot[lf.name]
except:
val_list_of_lists = taxon2value_list_barplot[int(lf.name)]
#col_count = 0
for col, value_list in enumerate(val_list_of_lists):
total = float(sum(value_list))
percentages = [(i / total) * 100 for i in value_list]
if col % 3 == 0:
col_list = colors2
else:
col_list = colors
b = StackedBarFace(percentages,
width=150,
height=18,
colors=col_list[0:len(percentages)])
b.rotation = 0
b.inner_border.color = "white"
b.inner_border.width = 0
b.margin_right = 5
b.margin_left = 5
if rotate:
b.rotation = 270
lf.add_face(b, col + col_add, position="aligned")
#col_count+=1
col_add += col + 1
if set2taxon2value_list_simple_barplot:
col_list = [
'#fc8d59', '#91bfdb', '#99d594', '#c51b7d', '#f1a340',
'#999999'
]
color_i = 0
col = 0
for one_set in header_list3:
if color_i > 5:
color_i = 0
color = col_list[color_i]
color_i += 1
# values for all taxons
values_lists = [
float(i) for i in
set2taxon2value_list_simple_barplot[one_set].values()
]
#print values_lists
#print one_set
value = set2taxon2value_list_simple_barplot[one_set][lf.name]
if set2taxon2value_list_simple_barplot_counts:
if isinstance(value, float):
a = TextFace(" %s " % str(round(value, 2)))
else:
a = TextFace(" %s " % str(value))
a.margin_top = 1
a.margin_right = 2
a.margin_left = 5
a.margin_bottom = 1
if rotate:
a.rotation = 270
lf.add_face(a, col + col_add, position="aligned")
#print 'value and max', value, max(values_lists)
fraction_biggest = (float(value) / max(values_lists)) * 100
fraction_rest = 100 - fraction_biggest
#print 'fractions', fraction_biggest, fraction_rest
b = StackedBarFace([fraction_biggest, fraction_rest],
width=100,
height=15,
colors=[color, 'white'])
b.rotation = 0
b.inner_border.color = "grey"
b.inner_border.width = 0
b.margin_right = 15
b.margin_left = 0
if rotate:
b.rotation = 270
if set2taxon2value_list_simple_barplot_counts:
if col == 0:
col += 1
lf.add_face(b, col + 1 + col_add, position="aligned")
col += 2
else:
lf.add_face(b, col + col_add, position="aligned")
col += 1
if set2taxon2value_list_simple_barplot_counts:
col_add += col
else:
col_add += col
if taxon2set2value_heatmap:
i = 0
#if not taxon2label:
# col_add-=1
for col2, head in enumerate(header_list2):
col_name = header_list2[i]
try:
value = taxon2set2value_heatmap[col_name][str(lf.name)]
except:
try:
value = taxon2set2value_heatmap[col_name][round(
float(lf.name), 2)]
except:
value = 0
if header_list2[i] == 'duplicates':
print('dupli', lf.name, value)
#print 'val----------------', value
if int(value) > 0:
if int(value) >= 10 and int(value) < 100:
n = TextFace('%4i' % value)
elif int(value) >= 100:
n = TextFace('%3i' % value)
else:
n = TextFace('%5i' % value)
n.margin_top = 1
n.margin_right = 2
n.margin_left = 5
n.margin_bottom = 1
n.hz_align = 1
n.vt_align = 1
if rotate:
n.rotation = 270
n.inner_background.color = rgb2hex(
column2scale[col_name][0].to_rgba(
float(value))) #"orange"
#print 'xaxaxaxaxa', value,
if float(value) > column2scale[col_name][1]:
n.fgcolor = 'white'
n.opacity = 1.
n.hz_align = 1
n.vt_align = 1
lf.add_face(n, col2 + col_add, position="aligned")
i += 1
else:
n = TextFace('')
n.margin_top = 1
n.margin_right = 1
n.margin_left = 5
n.margin_bottom = 1
n.inner_background.color = "white"
n.opacity = 1.
if rotate:
n.rotation = 270
lf.add_face(n, col2 + col_add, position="aligned")
i += 1
#lf.name = taxon2description[lf.name]
n = TextFace(taxon2description[lf.name],
fgcolor="black",
fsize=12,
fstyle='italic')
lf.add_face(n, 0)
for n in t1.traverse():
nstyle = NodeStyle()
if n.support < 1:
nstyle["fgcolor"] = "black"
nstyle["size"] = 6
n.set_style(nstyle)
else:
nstyle["fgcolor"] = "red"
nstyle["size"] = 0
n.set_style(nstyle)
return t1, tss
def plot_tree_text_metadata(tree_file, header2taxon2text, ordered_header_list,
biodb):
from chlamdb.biosqldb import manipulate_biosqldb
server, db = manipulate_biosqldb.load_db(biodb)
t1 = Tree(tree_file)
taxon2description = manipulate_biosqldb.taxon_id2genome_description(
server, biodb, filter_names=True)
# Calculate the midpoint node
R = t1.get_midpoint_outgroup()
# and set it as tree outgroup
t1.set_outgroup(R)
tss = TreeStyle()
tss.draw_guiding_lines = True
tss.guiding_lines_color = "gray"
tss.show_leaf_name = False
for i, leaf in enumerate(t1.iter_leaves()):
# first leaf, add headers
if i == 0:
for column, header in enumerate(ordered_header_list):
n = TextFace('%s' % (header))
n.margin_top = 0
n.margin_right = 1
n.margin_left = 20
n.margin_bottom = 1
n.rotation = 270
n.hz_align = 2
n.vt_align = 2
n.inner_background.color = "white"
n.opacity = 1.
tss.aligned_header.add_face(n, column)
for column, header in enumerate(ordered_header_list):
text = header2taxon2text[header][int(leaf.name)]
n = TextFace('%s' % text)
n.margin_top = 1
n.margin_right = 1
n.margin_left = 5
n.margin_bottom = 1
n.inner_background.color = "white"
n.opacity = 1.
#n.rotation = 270
leaf.add_face(n, column + 1, position="aligned")
# rename leaf (taxon_id => description)
n = TextFace(taxon2description[leaf.name],
fgcolor="black",
fsize=12,
fstyle='italic')
leaf.add_face(n, 0)
for n in t1.traverse():
# rename leaf
nstyle = NodeStyle()
if n.support < 1:
nstyle["fgcolor"] = "black"
nstyle["size"] = 6
n.set_style(nstyle)
else:
nstyle["fgcolor"] = "red"
nstyle["size"] = 0
n.set_style(nstyle)
return t1, tss
def plot_heat_tree_V1(taxid2n,
tree_file,
genes,
taxid2st=False,
leaf_label_conversion_dico=False):
'''
Plot heatmap next to a tree. The order of the heatmap **MUST** be the same,
as order of the leafs on the tree. The tree must be in the Newick format. If
*output_file* is specified, then heat-tree will be rendered as a PNG,
otherwise interactive browser will pop-up with your heat-tree.
TODO ajouter en option la possibilite d'ajouter en option la valeur dans la cellule
Parameters
----------
tree_file: str
Path to the tree file in Newick format. The leaf node labels should
be the same as as row names in the heatmap file. E.g. row1, row2.
output_file: str, optional
If specified the heat-tree will be rendered in that file as a PNG image,
otherwise interactive browser will pop-up. **N.B.** program will wait
for you to exit the browser before continuing.
'''
t1 = Tree(tree_file)
tss = TreeStyle()
#t.populate(8)
# Calculate the midpoint node
R = t1.get_midpoint_outgroup()
# and set it as tree outgroup
t1.set_outgroup(R) # To operate with numbers efficiently
import matplotlib.cm as cm
from matplotlib.colors import rgb2hex
import matplotlib as mpl
norm = mpl.colors.Normalize(vmin=0.8, vmax=1) # map2count[map[0]][0]
cmap_blue = cm.Blues
m2 = cm.ScalarMappable(norm=norm, cmap=cmap_blue)
leaf_number = 0
for lf in t1.iter_leaves():
leaf_number += 1
lf.branch_vertical_margin = 0
try:
data = taxid2n[str(lf.name)]
except:
data = [0]
try:
st = taxid2st[lf.name]
except:
st = False
'''
if "taxon2accession_list" not in locals():
from chlamdb.biosqldb import manipulate_biosqldb
server, db = manipulate_biosqldb.load_db("k_cosson_05_16")
sql = 'select taxon_id, accession from bioentry where biodatabase_id=104'
data_tax = server.adaptor.execute_and_fetchall(sql,)
taxon2accession_list = {}
for i in data_tax:
if i[0] not in taxon2accession_list:
taxon2accession_list[i[0]] = [i[1]]
else:
taxon2accession_list[i[0]].append(i[1])
else:
for taxon in taxon2accession_list:
if lf.name in taxon2accession_list[taxon]:
for accession in taxon2accession_list[taxon]:
print lf.name, accession
try:
st = taxid2st[accession]
data = taxid2n[accession]
print 'st ok!!', st
break
except:
continue
'''
if accession2description:
try:
lf.name = accession2description[lf.name]
except:
pass
if st:
lf.name = lf.name + ' (' + st + ')'
else:
pass
for col, value in enumerate(data):
if leaf_number == 1:
n = TextFace('%s' % (genes[col]), fsize=6)
n.vt_align = 2
n.hz_align = 2
n.rotation = 270
n.margin_top = 0
n.margin_right = 0
n.margin_left = 4
n.margin_bottom = 0
n.inner_background.color = "white"
n.opacity = 1.
tss.aligned_header.add_face(n, col)
#lf.add_face(n, col, position="aligned")
if value > 0:
n = TextFace(' ')
n.margin_top = 0
n.margin_right = 0
n.margin_left = 0
n.margin_bottom = 0
n.inner_background.color = rgb2hex(m2.to_rgba(
float(value))) #'#140718' #"#81BEF7"
n.opacity = 1.
lf.add_face(n, col, position="aligned")
else:
n = TextFace(' ')
n.margin_top = 0
n.margin_right = 0
n.margin_left = 0
n.margin_bottom = 0
n.inner_background.color = "white"
n.opacity = 1.
lf.add_face(n, col, position="aligned")
return t1, leaf_number, tss
def plot_tree_barplot(tree_file,
taxon2value_list_barplot,
header_list,
taxon2set2value_heatmap=False,
header_list2=False,
column_scale=True,
general_max=False,
barplot2percentage=False,
taxon2mlst=False):
'''
display one or more barplot
:param tree_file:
:param taxon2value_list:
:param exclude_outgroup:
:param bw_scale:
:param barplot2percentage: list of bool to indicates if the number are percentages and the range should be set to 0-100
:return:
'''
import matplotlib.cm as cm
from matplotlib.colors import rgb2hex
import matplotlib as mpl
if taxon2mlst:
mlst_list = list(set(taxon2mlst.values()))
mlst2color = dict(zip(mlst_list, get_spaced_colors(len(mlst_list))))
mlst2color['-'] = 'white'
if isinstance(tree_file, Tree):
t1 = tree_file
else:
t1 = Tree(tree_file)
# Calculate the midpoint node
R = t1.get_midpoint_outgroup()
# and set it as tree outgroup
t1.set_outgroup(R)
tss = TreeStyle()
value = 1
tss.draw_guiding_lines = True
tss.guiding_lines_color = "gray"
tss.show_leaf_name = False
if column_scale and header_list2:
import matplotlib.cm as cm
from matplotlib.colors import rgb2hex
import matplotlib as mpl
column2scale = {}
for column in header_list2:
values = taxon2set2value_heatmap[column].values()
norm = mpl.colors.Normalize(vmin=min(values), vmax=max(values))
cmap = cm.OrRd
m = cm.ScalarMappable(norm=norm, cmap=cmap)
column2scale[column] = m
cmap = cm.YlGnBu #YlOrRd#OrRd
values_lists = taxon2value_list_barplot.values()
scale_list = []
max_value_list = []
for n, header in enumerate(header_list):
#print 'scale', n, header
data = [float(i[n]) for i in values_lists]
if barplot2percentage is False:
max_value = max(data) #3424182#
min_value = min(data) #48.23
else:
if barplot2percentage[n] is True:
max_value = 100
min_value = 0
else:
max_value = max(data) #3424182#
min_value = min(data) #48.23
norm = mpl.colors.Normalize(vmin=min_value, vmax=max_value)
m1 = cm.ScalarMappable(norm=norm, cmap=cmap)
scale_list.append(m1)
if not general_max:
max_value_list.append(float(max_value))
else:
max_value_list.append(general_max)
for i, lf in enumerate(t1.iter_leaves()):
#if taxon2description[lf.name] == 'Pirellula staleyi DSM 6068':
# lf.name = 'Pirellula staleyi DSM 6068'
# continue
if i == 0:
col_add = 0
if taxon2mlst:
header_list = ['MLST'] + header_list
for col, header in enumerate(header_list):
#lf.add_face(n, column, position="aligned")
n = TextFace(' ')
n.margin_top = 1
n.margin_right = 2
n.margin_left = 2
n.margin_bottom = 1
n.rotation = 90
n.inner_background.color = "white"
n.opacity = 1.
n.hz_align = 2
n.vt_align = 2
tss.aligned_header.add_face(n, col_add + 1)
n = TextFace('%s' % header)
n.margin_top = 1
n.margin_right = 2
n.margin_left = 2
n.margin_bottom = 2
n.rotation = 270
n.inner_background.color = "white"
n.opacity = 1.
n.hz_align = 2
n.vt_align = 1
tss.aligned_header.add_face(n, col_add)
col_add += 2
if header_list2:
for col, header in enumerate(header_list2):
n = TextFace('%s' % header)
n.margin_top = 1
n.margin_right = 20
n.margin_left = 2
n.margin_bottom = 1
n.rotation = 270
n.hz_align = 2
n.vt_align = 2
n.inner_background.color = "white"
n.opacity = 1.
tss.aligned_header.add_face(n, col + col_add)
if taxon2mlst:
try:
#if lf.name in leaf2mlst or int(lf.name) in leaf2mlst:
n = TextFace(' %s ' % taxon2mlst[int(lf.name)])
n.inner_background.color = 'white'
m = TextFace(' ')
m.inner_background.color = mlst2color[taxon2mlst[int(lf.name)]]
except:
n = TextFace(' na ')
n.inner_background.color = "grey"
m = TextFace(' ')
m.inner_background.color = "white"
n.opacity = 1.
n.margin_top = 2
n.margin_right = 2
n.margin_left = 0
n.margin_bottom = 2
m.margin_top = 2
m.margin_right = 0
m.margin_left = 2
m.margin_bottom = 2
lf.add_face(m, 0, position="aligned")
lf.add_face(n, 1, position="aligned")
col_add = 2
else:
col_add = 0
try:
val_list = taxon2value_list_barplot[lf.name]
except:
if not taxon2mlst:
val_list = ['na'] * len(header_list)
else:
val_list = ['na'] * (len(header_list) - 1)
for col, value in enumerate(val_list):
# show value itself
try:
n = TextFace(' %s ' % str(value))
except:
n = TextFace(' %s ' % str(value))
n.margin_top = 1
n.margin_right = 5
n.margin_left = 10
n.margin_bottom = 1
n.inner_background.color = "white"
n.opacity = 1.
lf.add_face(n, col_add, position="aligned")
# show bar
try:
color = rgb2hex(scale_list[col].to_rgba(float(value)))
except:
color = 'white'
try:
percentage = (value / max_value_list[col]) * 100
#percentage = value
except:
percentage = 0
try:
maximum_bar = (
(max_value_list[col] - value) / max_value_list[col]) * 100
except:
maximum_bar = 0
#maximum_bar = 100-percentage
b = StackedBarFace([percentage, maximum_bar],
width=100,
height=10,
colors=[color, "white"])
b.rotation = 0
b.inner_border.color = "grey"
b.inner_border.width = 0
b.margin_right = 15
b.margin_left = 0
lf.add_face(b, col_add + 1, position="aligned")
col_add += 2
if taxon2set2value_heatmap:
shift = col + col_add + 1
i = 0
for col, col_name in enumerate(header_list2):
try:
value = taxon2set2value_heatmap[col_name][lf.name]
except:
try:
value = taxon2set2value_heatmap[col_name][int(lf.name)]
except:
value = 0
if int(value) > 0:
if int(value) > 9:
n = TextFace(' %i ' % int(value))
else:
n = TextFace(' %i ' % int(value))
n.margin_top = 1
n.margin_right = 1
n.margin_left = 20
n.margin_bottom = 1
n.fgcolor = "white"
n.inner_background.color = rgb2hex(
column2scale[col_name].to_rgba(
float(value))) #"orange"
n.opacity = 1.
lf.add_face(n, col + col_add, position="aligned")
i += 1
else:
n = TextFace(' ') #% str(value))
n.margin_top = 1
n.margin_right = 1
n.margin_left = 20
n.margin_bottom = 1
n.inner_background.color = "white"
n.opacity = 1.
lf.add_face(n, col + col_add, position="aligned")
n = TextFace(lf.name, fgcolor="black", fsize=12, fstyle='italic')
lf.add_face(n, 0)
for n in t1.traverse():
nstyle = NodeStyle()
if n.support < 1:
nstyle["fgcolor"] = "black"
nstyle["size"] = 6
n.set_style(nstyle)
else:
nstyle["fgcolor"] = "red"
nstyle["size"] = 0
n.set_style(nstyle)
return t1, tss
def heatmap_view(tree, orthologous_groups, save_dir):
"""Generates a heatmap of regulation states in all species."""
light_tree = copy.deepcopy(tree) # Tree copy for the light heatmap
# Heat map settings
rect_face_fgcolor = 'black'
locus_tag_len = max(len(gene.locus_tag) + 5
for ortho_grp in orthologous_groups
for gene in ortho_grp.genes)
rect_face_width = locus_tag_len * 8
light_rect_face_width = 20
rect_face_height = 20
rotation = 90
# Sort orthologous groups by the number of regulated genes in each group
orthologous_groups = filter_and_sort_orthologous_grps(orthologous_groups)
# For each species and its gene in each orthologous group, draw a rectangle
for node, light_node in zip(tree.get_leaves(), light_tree.get_leaves()):
for i, orthologous_grp in enumerate(orthologous_groups, start=1):
#get all orthologs in group
matching_genes = [g for g in orthologous_grp.genes \
if g.genome.strain_name == node.name]
#if there is ortholog
if len(matching_genes) > 0:
# Get the first ortholog from the genome in the group
#this is the one with higher probability of regulation.
#so this probability will be displayed for the group
gene = matching_genes[0]
p_regulation = gene.operon.regulation_probability
p_notregulation = 1.0 - p_regulation
p_absence = 0
# No ortholog from this genome
else:
gene = None
p_regulation = 0
p_notregulation = 0
p_absence = 1
# Color of the rectangle is based on probabilities
rect_face_bgcolor = rgb2hex(
p_notregulation, p_regulation, p_absence)
rect_face_text = ('%s [%d]' % (gene.locus_tag, gene.operon.operon_id)
if gene else '')
rect_face_label = {'text': rect_face_text,
'font': 'Courier',
'fontsize': 8,
'color': 'black'}
# Create the rectangle
rect_face = RectFace(rect_face_width, rect_face_height,
rect_face_fgcolor, rect_face_bgcolor,
label=rect_face_label)
light_rect_face = RectFace(light_rect_face_width, rect_face_height,
rect_face_fgcolor, rect_face_bgcolor,
label='')
rect_face.rotation = -rotation
light_rect_face.rotation = -rotation
# Add the rectangle to the corresponding column
node.add_face(rect_face, column=i, position='aligned')
light_node.add_face(light_rect_face, column=i, position='aligned')
ts = TreeStyle()
# Add orthologous group descriptions
descriptions = ['-'.join([grp.description, str(grp.NOGs)]) for grp in orthologous_groups]
max_description_len = max(map(len, descriptions))
descriptions = [
'[%d]' % i + description + ' '*(max_description_len-len(description))
for i, description in enumerate(descriptions, start=1)]
for i, description in enumerate(descriptions, start=1):
text_face = TextFace(description, ftype='Courier')
text_face.hz_align = 1
text_face.vt_align = 1
text_face.rotation = -rotation
ts.aligned_header.add_face(text_face, column=i)
# Rotate the generated heatmap.
ts.margin_left = 10
ts.margin_top = 20
ts.rotation = rotation
ts.show_scale = False
# For some reason, it can't render to PDF in color
tree.render(os.path.join(save_dir, 'heatmap.svg'), tree_style=ts)
light_tree.render(os.path.join(save_dir, 'heatmap_light.svg'), tree_style=ts)
