installation

You can build and install the binary using cargo

cargo install --git https://github.com/kopperud/splitfrequencies

usage

Call splitfrequencies --help for further info. For example

splitfrequencies --input primates_cytb*.trees --output splits.csv --burnin 0.1

If no output file is given, then the results are printed to stdout. Example:

splitfrequencies --input primates_cytb*.trees

split,primates_cytb_JC_run_1.trees,primates_cytb_JC_run_2.trees
[01111110111001110011110],0.001777382803821373,0.001777382803821373
[01100100000000000000000],0.8209286825149966,0.8682514996667408
[01000100000000000000000],0.8062652743834703,0.8589202399466785
[00000100000000010000000],0.17907131748500332,0.13174850033325927
...

algorithm for computing all splits:

Initial step: compute an ordered vector of all tip labels in the second line of the first file. Assume the tip labels are equal across all trees in all files.

sub-algorithm:

1. Begin at an internal node N.
2. Initialize a bit vector with zero elements, i.e. [0,0,0,0,...], representing presence or absence of the tip in the split.
3. Perform a traversal of the subtree descending from N. If a tip is visited, assign a 1 in the bit vector at the index representing the tip.
4. Store the bit vector representing the split

Repeat the above sub-algorithm for all internal nodes N, except the root. The left and right subtrees descending from the root will have complementary splits. For example, [1,1,0,0,0] for the left subtree, and [0,0,1,1,1] for the right subtree.

Trivial tip splits (i.e. [1,0,0,...]...) are not included. Symmetric or complementary are not always added, i.e. if the split [1,0,0,1] is recorded, it's not guaranteed that the split [0,1,1,0] is also included.