supertrees module in development

toytree/infer/src/supertrees.py

@@ -0,0 +1,58 @@
+#!/usr/bin/env python
+
+"""Implementatin of supertree inference methods.
+
+
+References
+----------
+https://pubmed.ncbi.nlm.nih.gov/39119174/
+"""
+
+import numpy as np
+from toytree.core import ToyTree
+from toytree.core.multitree import MultiTree
+
+
+def get_matrix_representation(trees: ToyTree, tiplabels: list[str]) -> np.ndarray:
+    """Return a matrix representation of a set of rooted trees.
+
+    This matrix can be analyzed using a phylogenetic inference tool
+    to infer a tree. A common approach is to infer a parsimony tree in
+    which case the supertree is termed a matrix representation with
+    parsimony (MRP) tree.
+    """
+    # get array filled w/ 1 for tips in each clade
+    data = []
+    for bipart in tree.enum.iter_bipartitions():
+        dat = np.zeros(len(tiplabels), dtype=int)
+        for i in bipart[0]:
+            dat[tiplabels.index(i)] = 1
+        data.append(dat)
+
+    # get mask of missing taxa in this tree
+    mask = np.zeros(len(tiplabels), dtype=np.bool_)
+    for i in set(tiplabels) - set(tree.get_tip_labels()):
+        mask[tiplabels.index(i)] = True
+
+    # create a masked array to
+    marr = np.ma.array(
+        data=data,
+        mask=[mask for i in data],
+        dtype=int,
+    )
+    print(marr.T)
+
+
+if __name__ == "__main__":
+
+    import toytree
+    t0 = "((a,b),((c,d),f));"
+    t1 = "((a,b),((c,d),e));"
+    t2 = "((a,c),((b,d),(e,f)));"
+    t3 = "((a,c),(b,e));"
+
+    names = list("abcdef")
+    trees = toytree.mtree([t0, t1, t2, t3])
+    for tree in trees:
+        marr = get_matrix_representation(tree, names)
+        print(marr)