Decomposition Into Single-COpy gene trees (DISCO) is a method for decomposing multi-copy gene-family trees while attempting to preserve orthologs and discard paralogs. These single-copy gene trees can be subsequently used by methods that can estimate species trees from single-copy gene trees such as ASTRAL or ASTRID in order to obtain an accurate estimation of the species tree. Additionally, DISCO can be paired with concatenation analysis with methods such as RAxML.
NOTE: For species tree estimation default settings are recommended; however, for orthology detection using
-m 2is recommended so small groups are retrieved. It is also highly recommended that you use the most recent version of DISCO, as it applies the latest bugfixes (see changelog).
If you use DISCO, please cite:
Willson, James, Mrinmoy Saha Roddur, Baqiao Liu, Paul Zaharias, and Tandy Warnow. “DISCO: species tree inference using multicopy gene family tree decomposition.” Systematic Biology 71, no. 3 (2022): 610-629.
bibtex
@article{willson2022disco,
title={DISCO: Species tree inference using multicopy gene family tree decomposition},
author={Willson, James and Roddur, Mrinmoy Saha and Liu, Baqiao and Zaharias, Paul and Warnow, Tandy},
journal={Systematic Biology},
volume={71},
number={3},
pages={610--629},
year={2022},
publisher={Oxford University Press}
}
Given a list of multi-copy gene trees, DISCO does the following for each tree:
- Root the tree and tag each internal vertex as either a duplication event or a speciation event in such a way that minimizes the total number of duplications and losses. We do this with the ASTRAL-Pro rooting and tagging algorithm (Zhang et. al. 2020).
- Decompose gene tree by splitting off the smallest subtree under every vertex tagged as a duplication from the leaves to the root until all duplication events are resolved; it returns the set of single-copy trees produced.
DISCO requires Python 3.12 or later and can be installed from PyPI:
pip install disco-gdlThis installs the disco command along with its dependencies (TreeSwift and Biopython).
The disco command has two subcommands: decomp, used to obtain single-copy trees through gene tree decomposition, and concat used to create concatenated alignments with the disco decomposition. The usage of these subcommands is outlined below. A help message can be displayed for either subcommand by running disco <command> -h.
Input: File containing list of multi-copy trees in newick format
Output: File containing resulting list of single-copy trees after decomposition in newick format
disco decomp -i <input_file> -o <output_file> -d <delimiter>
-i, --input Input newick tree file
-o, --output Output newick tree file
-d, --delimiter Delimiter separating species name from rest of leaf label
-s, --single-tree Output only single large tree per gene tree
-n, --nth-delimiter Split on Nth delimiter (only works with -d)
-m, --minimum Minimum number of taxa required for tree to be outputted
-v, --verbose Enable verbose output
--keep-labels Keep original leaf labels instead of using species name
--no-decomp Outputs rooted trees without decomposition
--outgroups Write outgroups (including ties) to txt file
--remove-in-paralogs Remove in-paralogs before rooting/scoring
disco decomp -i example/g_100-mult.treesInput: File containing list of multi-copy trees in newick format and set of alignment files corresponding to the gene families.
Output: Concatenated alignment file <output_prefix>-aln.<phy|fas> (and <output_prefix>-partitions.txt with -p)
disco concat -i <input_trees> -a <aln_list> -f <format> -o <output_prefix> -d <delimiter> -m <number>
It is necessary for the input newick trees given to disco concat to have unique leaf labels where the taxon name comes first and is separated from the rest of the name by some delimiter.
The -a argument should be given the path to an "alignment list" file containing the path to each alignment file you want to concatenate separated by a new line. The order of the alignment files is important---for each tree in the input newick tree file, there should be a corresponding alignment on the same line in the respective alignment list file.
-i, --input Input newick tree file
-a, --alignment Text file containing paths to alignment files
-f, --format Format of alignment file (either "fasta" or "phylip")
-o, --output-prefix Prefix for output files
-m, --filter Minimum number of taxa required sequence group to be included
-d, --delimiter Delimiter separating species name from rest of leaf label
-p, --partition Create partition file
disco concat -i example/g_100.trees -o example -a example/seq_list.csv -f phylipdisco concat now has the option to create a partition file; this can be done with the optional -p argument. If this argument is specified, it is necessary to provide relevant information in the alignment list file. For example, see example/seq_list.csv. This would generate a partition file like so:
GTR+G, 0001=1-200
GTR+G, 0002=201-700
GTR+G, 0004=701-1300
...
Note: Gene 0003 is missing. This is not a mistake; this is because it does not have enough informative information to generate large enough DISCO subtrees and thus is not included in the alignment.