BSC5936-Fall 2005-PB,FR Computational Evolutionary Biology
Assignment 4
1. Write a program that uses stepwise addition to heuristically find the best tree for the full
12-taxon data set that we have been working with since Assignment 2. You can use an
arbitrary ordering of tips and then use your exhaustive enumeration algorithm with a few
modifications to generate all the trees with n+ 1 tips from a tree with ntips. At each step
of the algorithm, calculate the length of all trees with one extra tip that are possible to build
from your starting tree. Then choose the best one and use that one for the next step in the
algorithm. Use Fitch parsimony when you calculate tree scores.
2. Modify your program so that it can hold a user-specified number of trees at each step in the
addition process. For instance, if the user wants to save three trees, then you first construct
all the possible four-tip trees (three). Now you take the first tree and add the next tip to all
possible positions, calculating the score for each tree. Then you save the three best trees and
continue with the next starting tree. Again you add the next tip in all possible positions and
replace the worst of the three trees in memory every time you find a tree with a better score.
At the end of the search, go through all trees in memory and delete the ones that are longer
than the best tree(s).
3. (Optional) Program a branch swapping algorithm of your choice to improve the best tree
resulting from your stepwise addition algorithm. We recommend that you program either
NNI or the TBR algorithm with a neighborhood variable that determines how far away from
the original position (in number of nodes) that you will move the attachment point of either
subtree on the other subtree.
1
