AGO is a tool aimed at creating pipelines of existing bioinformatics tools for reconstructing ancestral syntenies and gene orders from extant gene orders and homologous gene families, in a phylogenetic context described by a rooted species tree.
AGO is based on the general methodology described in Duplication, Rearrangement and Reconciliation: A Follow-Up 13 Years Later (see also the review Comparative Methods for Reconstructing Ancient Genome Organization). It consists of the following steps:
- computing a multiple sequences alignment (MSA) for each homologous gene family;
- computing a gene tree, or sample of gene trees, from the MSA for each gene family;
- computing a reconciled gene tree, from the MSA and/or gene tree(s), for each gene family;
- computing ancestral gene adjacency candidates from the species tree, the extant gene orders and the reconciled gene trees;
- clearing conflicts from the set of candidate ancestral gene adjacencies (variant of the Small Parsimony Problem for gene orders).
This approach allows to account for the full complement of genes in extant genomes, at the expense of necessitating several steps, implemented in various tools, that are not trivial to interface together into a pipeline. AGO is aimed at easing the creation of such pipelines; AGO in itself does not introduce any novel algorithm, and aims solely at allowing to create pipelines based on published existing tools.
AGO has been developed using python version 3.8.10. We recommend to run AGO within a python virtual environment (https://docs.python.org/3.8/library/venv.html) including the following python packages, that are required:
- ete3, version 3.1.2
- matplotlib, version 3.7.0.
- networkx, version 3.1.
- numpy, version 1.24.2.
- pandas, version 1.5.3.
- PyYAML, version 6.0.
- Biopython, version 1.81.
AGO allows to include the following external tools in an ancestral gene orders reconstruction pipeline:
- Multiple sequence alignments: MACSE, version 2.06; requires Java, OpenJDK version 17.0.2.
- Gene trees: IQ-TREE, version 2.0.7.
- Reconciled gene trees: ALE, version 1.0.
- Reconciled gene trees: GeneRax, version 2.1.0.
- Ancestral adjacencies: DeCoSTAR, version 1.2.
- Ancestral gene orders: SPP_DCJ; requires gurobi, version 10.0.1.
These tools require to be installed independently from AGO; we refer to the specific documentation of each tool for the installation instructions; a future release of AGO will include these tools in a Singularity container.
Finally AGO can run the external tools integrated into a pipeline either on local workstations (through shell scripts) or on High-Performance Computing (HPC) clusters using the slurm scheduler.
The AGO manual describes how to create and run a pipeline.
AGO is illustrated in a test case, focusing on the reconstruction of the ancestral gene orders of the X chromosome of Anopheles mosquitoes genomes using different implementations of the methodology described above:
- Computing reconciled gene trees from MSAs using GeneRax.
- Computing gene trees from MSAs using IQ-TREE and reconciled gene trees from the gene trees using ALE.
- Computing gene trees from MSAs using IQ-TREE and reconciled gene trees using DeCoSTAR.
These experiments are described in example/README.md.
Authors: Evan Cribbie, Daniel Doerr, Cedric Chauve (cedric.chauve@sfu.ca).