A sample script using the ETE package to detect shift in internal nodes where a significant change in enrichment values (could be anything) happens

parsetree.py

#A sample script using the ETE package to detect shift in internal nodes where a significant change in enrichment values (could be anything) happens



#An example (using the scipy python module to perform a K-S test):

from scipy import stats
from ete2 import Tree

newick = "((((A, B)edge1, C)edge2, ((D, E)edge3, F)edge4)edge5, (((G, H)edge6, I)edge7, ((J, K)edge8, L)edge9)edge10)RootEdge;"

enrichment = {
    "A": 10,
    "B": 11,
    "C": 12,
    "D": 40,
    "E": 41,
    "F": 42,
    "G": 91,
    "H": 92,
    "I": 90,
    "J": 92,
    "K": 91,
    "L": 95 }

# load in format 1 to keep edge names in newick
t = Tree(newick, format=1)
print "Tree loaded"
print t.get_ascii()

# load enrichment info into the tree
for n in t.traverse():
    if n.is_leaf():
        n.add_feature("enrich", enrichment[n.name])

print "Test starting..."

# Analyze the tree
for n in t.traverse():
    if not n.is_leaf():
        # list of all  enrichment values in one branch
        e1 = [leaf.enrich for leaf in n.children[0].get_leaves()]
        # list of all  enrichment values in the sister branch
        e2 = [leaf.enrich for leaf in n.children[1].get_leaves()]

        # I use scipy.stats to perform ks two sample test. Change this
        # to the most appropriate method for your data.
        D, pvalue = stats.ks_2samp(e1, e2)
        #print pvalue, D
        if pvalue < 0.05:
            print "significant shift found in ", n.name
            print "pvalue =", pvalue
            print "enrichment one side", e1
            print "enrichment sister branch", e2
            print n

Produces:

Tree loaded

                    /-A
               /edge1
          /edge2    \-B
         |    |
         |     \-C
     /edge5
    |    |          /-D
    |    |     /edge3
    |     \edge4    \-E
    |         |
    |          \-F
-RootEdge
    |               /-G
    |          /edge6
    |     /edge7    \-H
    |    |    |
    |    |     \-I
     \edge10
         |          /-J
         |     /edge8
          \edge9    \-K
              |
               \-L
Test starting...
significant shift found in  RootEdge
pvalue = 0.00129974398539
enrichment one side [12, 42, 10, 11, 40, 41]
enrichment sister branch [90, 95, 91, 92, 92, 91]

                    /-A
               /---|
          /---|     \-B
         |    |
         |     \-C
     /---|
    |    |          /-D
    |    |     /---|
    |     \---|     \-E
    |         |
    |          \-F
----|
    |               /-G
    |          /---|
    |     /---|     \-H
    |    |    |
    |    |     \-I
     \---|
         |          /-J
         |     /---|
          \---|     \-K
              |
               \-L
significant shift found in  edge5
pvalue = 0.032621651652
enrichment one side [12, 10, 11]
enrichment sister branch [42, 40, 41]

               /-A
          /---|
     /---|     \-B
    |    |
    |     \-C
----|
    |          /-D
    |     /---|
     \---|     \-E
         |
          \-F