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)
# |
# | /-F
# \--------|
# | /-G
# \--------|
# \-H
# Now we can ask the numerical profile associated to each node
A = t.search_nodes(name='A')[0]
print "A associated profile:\n", A.profile
# [-1.23 -0.81 1.79 0.78 -0.42 -0.69 0.58]
#
# Or we can ask for the mean numerical profile of an internal
# partition, which is computed as the average of all vectors under the
# the given node.
cluster = t.get_common_ancestor("E", "A")
print "Internal cluster mean profile:\n", cluster.profile
#[-1.574 -0.686 1.048 -0.012 -0.118 0.614 0.728]
#
# We can also obtain the std. deviation vector of the mean profile
print "Internal cluster std deviation profile:\n", cluster.deviation
#[ 0.36565558 0.41301816 0.40676283 0.56211743 0.50704635 0.94949671
# 0.26753691]
# If would need to re-link the tree to a different matrix or use
# different matrix for different sub parts of the tree, we can use the
# link_to_arraytable method()
#
# Creates a matrix with all values = 1
matrix_ones = """
#Names\tcol1\tcol2\tcol3\tcol4\tcol5\tcol6\tcol7
A\t1\t1\t1\t1\t1\t1\t1
# |
# | /-F
# \--------|
# | /-G
# \--------|
# \-H
# Now we can ask the numerical profile associated to each node
A = t.search_nodes(name='A')[0]
print "A associated profile:\n", A.profile
# [-1.23 -0.81 1.79 0.78 -0.42 -0.69 0.58]
#
# Or we can ask for the mean numerical profile of an internal
# partition, which is computed as the average of all vectors under the
# the given node.
cluster = t.get_common_ancestor("E", "A")
print "Internal cluster mean profile:\n", cluster.profile
#[-1.574 -0.686 1.048 -0.012 -0.118 0.614 0.728]
#
# We can also obtain the std. deviation vector of the mean profile
print "Internal cluster std deviation profile:\n", cluster.deviation
#[ 0.36565558 0.41301816 0.40676283 0.56211743 0.50704635 0.94949671
# 0.26753691]
# If would need to re-link the tree to a different matrix or use
# different matrix for different sub parts of the tree, we can use the
# link_to_arraytable method()
#
# Creates a matrix with all values = 1
matrix_ones = """
#Names\tcol1\tcol2\tcol3\tcol4\tcol5\tcol6\tcol7
A\t1\t1\t1\t1\t1\t1\t1
