CAMUS (Constrained Algorithm Maximizing qUartetS) is a dynamic programming algorithm for inferring level-1 phylogenetic networks from quartets and a constraint tree.

If you use CAMUS, please cite

Willson, James, and Tandy Warnow. "CAMUS: Scalable Phylogenetic Network Estimation." bioRxiv (2026): 2026-02.

Given a rooted, binary, constraint tree $T$ and a list of input trees $\mathcal{G}$, the CAMUS algorithm finds the level-1 networks inducing the maximum number of quartets from $\mathcal{G}$ that contains the constraint tree $T$

CAMUS finds networks for different values of $k$, where $k$ is the number of edges added to the network. For example, if the best possible network that contains $T$ has $m$ edges, then CAMUS will output $m$ networks $N_1, N_2, \cdots, N_k, \cdots, N_m$, where $N_k$ is optimal under the constraint that it contains exactly $k$ edges.

CAMUS has the following inputs and outputs:

• Input

  • Constraint Tree: Rooted, binary, tree in newick format without duplicate labels.
  • Gene Trees: List of trees in newick format, containing only labels from the constraint tree.

• Output

  • Output Network: Level-1 networks written in extended newick format.

CAMUS should be invoked with the constraint tree file path and gene trees file path as positional arguments in that order; the output network and logging information is written to files with a prefix that can optionally be set with the -o flag.

CAMUS should be able to build on any operating system, though it has only been tested on Linux. It is implemented in the Go programming language (which can be installed from the Go website).

There are three methods for installing CAMUS.

After installing Go, run:

go install github.com/jsdoublel/camus@latest

This will install the CAMUS executable to $GOPATH/bin. You can check where that is using the command go env GOPATH; if you want to run CAMUS from any directory make sure you add $GOPATH/bin to your PATH environmental variable.

Once Go is installed, clone the repository and build the project with the following commands:

git clone https://github.com/jsdoublel/camus.git
cd camus
go build

The GitHub releases contain binaries compiled for Windows, macOS, and Linux (for both arm64 and amd64).

camus [ -f <format> | -o <output> | -t <threshold> | -n <threads> | -h | -v | ... ] <const_tree> <gene_trees>

There are two positional arguments indicating the inputs. Additionally, there are the following flags that precedes the positional arguments.

• Basic Flags

  • -f format [ newick | nexus ] (default "newick") sets the format of the input gene tree file
  • -t threshold [0, 1] (default 0.5) quartet filtering threshold
  • -n num_procs number of parallel processes
  • -o prefix output prefix
  • -s threshold collapse edges in gene trees with support less than threshold value
  • -h prints usage information and exits
  • -hh prints extended usage information and exits
  • -v prints software version and exits

• Experimental Flags

  • -sm mode [ max | norm | sym ] (default "max") sets the score mode
  • -a alpha parameter that adjusts penalty in ``sym" score mode
  • -asSet quartet count is calculated as a set (counts total unique quartet topologies)
  • -q mode [0, 2] (default 0) quartet filtering mode

Quartet filtering mode filters out less frequent quartet topologies. Mode -q 0 disables quartet filtering; -q 1 applies a less restrictive quartet filtering, and -q 2 is the most restrictive and recommended quartet filtering.

Score Modes are various modifications to the optimization score beyond simple maximization. These are experimental and maximization is recommended.