collapse tree nodes if leafs does not contain a specific feature

Question

[Abdullah]

Hi,

I have a random tree where I need to show only the interesting parts.

Each node in the tree contain a feature called feat, it only contains a value for some leafs.

What I want to do is:

Draw the tree, and only keep the parts where at least one leaf contain a value in this feature, i.e. collapse all the internal nodes where there are no leafs containing a value inside the feature feat.

I tried to use the collapsing functions used in the ETE2 documentation, but they don't seem to work well and they produce a non-complete newick tree which is not readable by the ete2 module (or maybe I did something wrong).

Can someone tell me how to do that?

Answer 1

[jhc]

I do similar things with ETE very often. Example:

from ete2 import Tree, TreeStyle
import random

def collapse_layout(node):
    # if none of the leaves in a given partition have at least a feat attribute
    # set to 1, collapse the node
    if not node.is_leaf() and 1 not in NODE2FEAT_CONTENT[node]:
        node.img_style["draw_descendants"] = False
        node.img_style["size"] = 20
        node.name = "Collapsed leaves"

t = Tree()
# generate a random tree and feat values
t.populate(100)
for leaf in t:
    leaf.add_features(feat=random.randint(0, 1))

# a cached dictionary informing about the feat attribute content for all nodes
NODE2FEAT_CONTENT = t.get_cached_content(store_attr="feat")

ts = TreeStyle()
ts.layout_fn = collapse_layout
t.show(tree_style=ts)