Python versus Julia distance counts

[Colin Beckingham]

The website rosalind.info offers a puzzle where we are given a set of networks in very loose Newick format and asked to find the distance between two specific taxa for each. A popular way to do this is to use the Phylo import from Bio into Python. Since I am learning about these networks I decided to use PhyloNetworks with Julia as my route to the solution. I was successful in massaging the loose Newick trees into a form that PhyloNetworks would accept and submitted an answer to a random set of about 40 networks and received a result that my answer was incorrect.

So I ran a Python routine on the same dataset to see what Python would make of the puzzle. Indeed it came up with a slightly different answer. In two puzzles I tried with a set of about 40 nets, 3 of the 40 counts were different. Different by exactly 2 more than the Python version.

I just mention this as a point of interest. Evidently I introduced a slight bump into my version of the net after massaging that caused the count to increase, or there is a difference in the counting in Python Bio and Phylonetworks.jl.