UniFrac implememtation in Rust

This is an example repo to show how to compute the UniFrac distance between a pair of samples containing taxa. It uses the phylotree crate to parse the tree file and feature-sample table (OTU table for example) and then compute the unweighted unifrac distance. This is the example implementation, not heavily optimized.

This is the orignal unweighted UniFrac, which is normalized by tree size for each pair of samples. That is the tree is trimmed first to include features in either of the sample or both but remove features that are not there for any of the samples. This unweighted is not a metric distance since the total space for each pair of samples is different.

Install

git clone https://github.com/jianshu93/unifrac-rs-test
cd unifrac-rs-test
cargo build --release
./target/release/unifrac -h

Usage

 ************** initializing logger *****************

Fast Unweighted UniFrac

Usage: unifrac --tree <TREE_FILE> --input <TABLE_FILE> --output <OUTPUT_FILE>

Options:
  -t, --tree <TREE_FILE>      Input newick format tree file
  -i, --input <TABLE_FILE>    Input tab-delimited sample-feature table
  -o, --output <OUTPUT_FILE>  Output file for distance matrix
  -h, --help                  Print help
  -V, --version               Print version

example

### remove bootstrap support first if you have it

### Then run unifrac like this:
unifrac -t data/test_rot_new2.nwk -i data/table.txt -o try.txt
cat try.txt

References

1.Lozupone, C. and Knight, R., 2005. UniFrac: a new phylogenetic method for comparing microbial communities. Applied and environmental microbiology, 71(12), pp.8228-8235.

2.Hamady, M., Lozupone, C. and Knight, R., 2010. Fast UniFrac: facilitating high-throughput phylogenetic analyses of microbial communities including analysis of pyrosequencing and PhyloChip data. The ISME journal, 4(1), pp.17-27.