1

I'm very new to DendroPy. What I want to do seems simple but I can't figure how to do it correctly and I didn't find anything on the internet.

I want to add a node midway between two nodes in a existing rooted dendropy tree.

from dendropy import Tree, Taxon, Node

t1 = Tree.get_from_string(
    "(((sp1: 0.35, sp2: 0.15):0.75, sp3:1): 0.5, (sp4: 0.5, sp5: 0.05)MRCA_sp4&sp5: 1)root;",
    "newick", rooting='force-rooted'
)
t1.print_plot()
mrca = t1.mrca(taxon_labels=["sp4", "sp5"])
print(mrca.description())

Tree and MRCA node description

The MRCA of sp4 and sp5 is found correctly. Now I'm trying to add a node midway between MRCA and root, using the code below:

def add_node_midway_between_2_nodes(lowernode, taxon_label=None, node_label=None):
    newtaxon = Taxon(label=taxon_label)
    newnode = Node(taxon=newtaxon, label=node_label)
    newnode.parent_node = lowernode.parent_node
    newnode.edge_length = lowernode.edge_length/2
    lowernode.parent_node = newnode
    lowernode.edge_length = newnode.edge_length
    return newnode

node = add_node_midway_between_2_nodes(mrca, node_label="midway between root and MRCA sp4&sp5")
t1.print_plot()
str_t1 = t1.as_string(schema='newick')
print(str_t1)

Tree with a node between root and MRCA sp4&sp5

[&R] (((sp1:0.35,sp2:0.15):0.75,sp3:1.0):0.5,((sp4:0.5,sp5:0.05)MRCA_sp4&sp5:0.5)midway_between_root_and_MRCA_sp4&sp5:0.5)root;

Looking at the plot and at the string it seems to have worked. But then when I try to compute the MRCA of sp4 et sp5 again, it doesn't find "MRCA sp4&sp5" anymore but the root node.

mrca = t1.mrca(taxon_labels=["sp4", "sp5"])
print(mrca.description())

Output = Description of root node

Going through parent_node from sp5, I do still find "MRCA sp4&sp5".

Out of desperation, I tried to redo the tree using the string str_t1, but it doesn't work either and even gives me another result (just as incorrect): the node "midway between root and MRCA sp4&sp5"

t1 = Tree.get_from_string(
    str_t1,
    "newick", rooting='force-rooted'
)
mrca = t1.mrca(taxon_labels=["sp4", "sp5"])
print(mrca.description())

Output = description of node "midway between root and MRCA sp4&sp5"

So what is a clean way to add a node midway between two nodes, that doesn't create weird events afterward?

Thank you very much

Vovin
  • 720
  • 4
  • 16
ALB
  • 13
  • 2

1 Answers1

0

Your code is mostly working but you should update_taxon_namespace and update_bipartitions to apply any changes of a tree topology correctly as it was recommended in documentation. So, in your case, it would look like this:

def add_node_midway_between_2_nodes(lowernode, taxon_label=None, node_label=None):
    newtaxon = Taxon(label=taxon_label)
    newnode = Node(taxon=newtaxon, label=node_label)
    newnode.parent_node = lowernode.parent_node
    newnode.edge_length = lowernode.edge_length/2
    lowernode.parent_node = newnode
    lowernode.edge_length = newnode.edge_length
    return newnode

node = add_node_midway_between_2_nodes(
    mrca, node_label="midway between root and MRCA sp4&sp5"
)
t1.update_taxon_namespace()
t1.update_bipartitions(
    suppress_unifurcations=False, suppress_storage=True
)  # suppress_storage is optional, I just do not want to create a bipartitions list
t1.print_plot()
str_t1 = t1.as_string(schema='newick')
print(str_t1)

NB!

Updating taxa namespace should be prior to updating bipartitions as the latter has to use a correct TaxonNamespace. Otherwise, you still get the strange behavior.

Howbeit, it is better to use builtin Node methods for fine tree reconstruction. For instance, I would rewrite the function such way:

def insert_new_node_posterior(
    node: Node,
    *,
    taxon_label: Optional[str] = None,
    node_label: Optional[str] = None,
    edge_length: Real,
    # If it was product or at least reusable in the future code,
    # I would add more arguments for proportion specification,
    # using height, distance from root &c.
) -> Node:
    parent = node.parent_node
    if not parent:
        raise Exception("You cannot insert a node in posterior to the root.")

    new_taxon = Taxon(label=taxon_label)
    i = parent.child_nodes().index(node)
    parent.remove_child(node)
    intermediate_node = parent.insert_new_child(
        index=i, taxon=new_taxon, label=node_label, edge_length=edge_length
    )
    node.edge_length -= edge_length
    intermediate_node.add_child(node)
    return intermediate_node


node = insert_new_node_posterior(
    mrca,
    node_label="midway between root and MRCA sp4&sp5",
    edge_length=mrca.edge_length / 2
)
t1.update_taxon_namespace()
t1.update_bipartitions(
    suppress_unifurcations=False, suppress_storage=True
)

t1.print_plot()
str_t1 = t1.as_string(schema='newick')
print(str_t1)

Nonetheless, Tree.mrca still shows an improper node:

Node object at 0x1be8b959460<Node object at 0x1be8b959460: 'midway between root and MRCA sp4&sp5' (<Taxon 0x1be8cdceeb0 'None'>)>
    [Edge]
        Edge object at 0x1be8b959400 (1917897249792, Length=0.5)
    [Taxon]
         Taxon object at 0x1be8cdceeb0: <Unnamed Taxon>
    [Parent]
        Node object at 0x1be8c3b2b80<Node object at 0x1be8c3b2b80: 'root' (None)>
    [Children]
        [0] Node object at 0x1be8be4d6a0<Node object at 0x1be8be4d6a0: 'MRCA sp4&sp5' (None)>

Although, this is not a bug in the case as this is just a feature of the method. Due to this in the source code:

if cms:
    # for at least one taxon cm has 1 and bipartition has 1
    if cms == leafset_bitmask:
        # curr_node has all of the 1's that bipartition has
        if cm == leafset_bitmask:
            return curr_node  # Vovin's comment: Since there is a unifurcation,
                              # it returns the current node
                              # instead of the next iteration
        last_match = curr_node
        nd_source = iter(curr_node.child_nodes())
    else:
        # we have reached a child that has some, but not all of the
        #   required taxa as descendants, so we return the last_match
        return last_match

For example, if we add a child to the new node, it works well:

node = insert_new_node_posterior(
    mrca,
    node_label="midway between root and MRCA sp4&sp5",
    edge_length=mrca.edge_length / 2
)
node.new_child(label="sp6", edge_length=1, taxon=Taxon(label="sp6"))
t1.update_taxon_namespace()
t1.update_bipartitions(
    suppress_unifurcations=False, suppress_storage=True
)
mrca = t1.mrca(taxon_labels=["sp4", "sp5"])
print(mrca.description())

Node object at 0x1be8c1e22b0<Node object at 0x1be8c1e22b0: 'MRCA sp4&sp5' (None)>
    [Edge]
        Edge object at 0x1be8c1e2340 (1917906199360, Length=0.5)
    [Taxon]
        None
    [Parent]
        Node object at 0x1be8ba8da30<Node object at 0x1be8ba8da30: 'midway between root and MRCA sp4&sp5' (<Taxon 0x1be8ba8d370 'None'>)>
    [Children]
        [0] Node object at 0x1be8c1e2910<Node object at 0x1be8c1e2910: 'None' (<Taxon 0x1be8c1e2640 'sp4'>)>
        [1] Node object at 0x1be8c1e28e0<Node object at 0x1be8c1e28e0: 'None' (<Taxon 0x1be8c1e2e50 'sp5'>)>
Vovin
  • 720
  • 4
  • 16