One Permutation Hashing with Optimal Densification (1,2) can be use for genomic distance estimation (1-ANI) and then we can perform rapid neighbor-joining (4) based on the genomic distance. We also provided a new densification strategy called faster densification (or reverse optimal densification) (3), which is more accurate and faster for large sketch size (5). Bindasthree is more than 1000 times faster than alternatives, e.g., mashtree while providing similar accuracy. We can run bindashtree for 20,000 E. Coli genomes in just a few minutes with multiple threads (e.g., 64).
conda install -c bioconda -c conda-forge bindashtree
brew tap jianshu93/bindashtree
brew update
brew install bindashtree### Install from cargo, install Rustup first here: https://rustup.rs, cargo will be installed by default
cargo install bindashtree### Pre-built library (Linux only)
https://github.com/jianshu93/bindashtree/releases/download/v0.1.1/bindashtree_Linux_x86-64_v0.1.1.zip
unzip bindashtree_Linux_86-64_v0.1.0.zip
chmod a+x ./bindashtree
./bindashtree -h### compling from source
git clone https://github.com/jianshu93/bindashtree.git
cd bindashtree
cargo build --release
./target/release/bindashtree -h ************** initializing logger *****************
Binwise Densified MinHash and Rapid Neighbor-joining Tree Construction
Usage: bindashtree [OPTIONS] --input <INPUT_LIST_FILE> --output_tree <OUTPUT_TREE_FILE>
Options:
-i, --input <INPUT_LIST_FILE>
Genome list file (one FASTA/FNA file per line), gz supported
-k, --kmer_size <KMER_SIZE>
K-mer size [default: 16]
-s, --sketch_size <SKETCH_SIZE>
MinHash sketch size [default: 10240]
-d, --densification <DENS_OPT>
Densification strategy: 0=Optimal Densification, 1=Reverse Optimal Densification/faster Densification [default: 0]
-t, --threads <THREADS>
Number of threads to use in parallel [default: 1]
--tree <TREE_METHOD>
Tree construction method: naive, rapidnj, hybrid [default: rapidnj]
--chunk_size <chunk_size>
Chunk size for RapidNJ/Hybrid methods [default: 30]
--naive_percentage <naive_percentage>
Percentage of steps naive for hybrid method [default: 90]
--output_matrix <OUTPUT_MATRIX_FILE>
Output the phylip distance matrix to a file
--output_tree <OUTPUT_TREE_FILE>
Output the resulting tree in Newick format to a file
-h, --help
Print help
-V, --version
Print version
A newick format tree and phylip format distance matrix will be the output depending on your options. Tree can be visualized via Figtree, iTOL or ggtree
ls ./data/*.fna.gz > name.txt
#### for highly similar genomes, e.g., > 99.9% ANI, a large sketch size should be used. -s 10204 works well for ANI below 99%.
./target/release/bindashtree -i name.txt -k 16 -s 10240 -d 1 -t 8 --output_tree try.nwk1.Li, P., Owen, A. and Zhang, C.H., 2012. One permutation hashing. Advances in Neural Information Processing Systems, 25.
2.Shrivastava, A., 2017, July. Optimal densification for fast and accurate minwise hashing. In International Conference on Machine Learning (pp. 3154-3163). PMLR.
3.Mai, T., Rao, A., Kapilevich, M., Rossi, R., Abbasi-Yadkori, Y. and Sinha, R., 2020, August. On densification for minwise hashing. In Uncertainty in Artificial Intelligence (pp. 831-840). PMLR.
4.Simonsen, M., Mailund, T. and Pedersen, C.N., 2008. Rapid neighbour-joining. In Algorithms in Bioinformatics: 8th International Workshop, WABI 2008, Karlsruhe, Germany, September 15-19, 2008. Proceedings 8 (pp. 113-122). Springer Berlin Heidelberg.
5.Zhao, J., Zhao, X., Pierre-Both, J. and Konstantinidis, K.T., 2024. BinDash 2.0: new MinHash scheme allows ultra-fast and accurate genome search and comparisons. bioRxiv, pp.2024-03.