This package fits various models of DNA evolution to a collection of two-taxon trees, each using data from a single pairwise alignment.
In contrast to standard phylogenetic inference programs, fit-star allows the root state of each 2-tree to be fixed to the reference sequence within the pairwise alignment.
Additionally, it assumes that each sequence evolved independently from the known start state - along a star-like phylogeny.
This is unusual. For inferring full phylogenies RAxML, PhyML, or FastTree will be more useful.
An application to B cell receptor sequences is described in our preprint on arXiv.
- libprotobuf-2.4.1-dev
- beagle-2.1
- bpp-core-dev 2.1.0
- bpp-phyl-dev 2.1.0
- bpp-seq-dev 2.1.0
- libz-dev
- eigen (included)
- nlopt (included)
- samtools (included)
- boost (included)
Functionality is separated into two programs:
build-mutation-matrices transforms each aligned sequence in a BAM file to a 4x4 matrix containing counts of each nucleotide substitution.
If multiple alignments are available for a single sequence, and the sequences are sorted by name, each alignment can form a partition to be fit by fit-star.
fit-star fits various evolutionary models (TN93, HKY85, GTR) to the pairwise substitution matrices generated by build-mutation-matrices.
Both model parameters and branch lengths are optimized.
Optimized parameter values and branch lengths are output as JSON.
fit-star is multithreaded using OpenMP, and by default uses all the cores available on your machine.
To reduce the number of cores used, set the environment variable OMP_NUM_THREADS.
Assuming all required dependencies are present, just run make.
Binaries will be placed in build/src/.
Assuming all required dependencies are present, just run make test.