NSB is a tool for alignment-free phylogenetic distance estimation, under a no strand-bias, time reversible GTR model, TK4.
Alignment-free methods are useful because of their simplification of the pipeline/process of phylogenetic inference. However, despite the appeal, the accuracies of the alignment-based methods most often surpass the ones with an alignment-free setting, as the alignment-free methods use simple base-substitution GTR models, like Jukes-Cantor.
NSB uses a base-substitution technique on k-mers to identify the frequencies of transitions and transversions, and thus allows the use of more complex sequence evaluation models. This enables NSB to estimate more accurate phylogenetic distances, even when the true distances are high.
- Python 3
- NumPy, Pandas, Pickle
- Skmer
You can install NSB using the installation script install.sh we provide. It creates a new Anaconda environment called nsb. After environment is created, active it using the following command:
conda activate nsb
- NSB takes input in fasta, fastq and fna formats.
- All the sequence files are needed to be saved in a directory, i.e. see ref_dir folder for example.
The output to NSB is a n*n distance matrix, saved under the name ref-dist-mat-nsb-ref_dir.txt.
python NSB_jf_free.py reference -m 2 -k 31 -s 100000 -p 4 ref_dir