EGTool: an evolutionary genome topology tool for multiscale comparative genomics
Fatih Sarigol, Darrin T Schultz, Oleg SimakovAbstract
Summary
Increased phylogenomic sampling and availability of thousands of high-quality genome assemblies require tools to simultaneously analyze changes to both chromosomal and sub-chromosomal genome architecture. Recently, we proposed an Evolutionary Genome Topology (EGT) approach to identify and associate genome structural changes with evolutionary transitions. In this Application Note, we provide a generalized implementation of this approach, which we call EGTool. EGTool compares genomes to each other (multigenome topology) or loci to each other across genomes (multilocus topology), accepts arbitrary positional and orthology inputs, works in base-pair or index coordinates, projects new genomes onto an existing embedding without refitting, and recovers the features that characterize any chosen region of the projection via inverse transform. A 510-genome, 1,274 locus example is included to illustrate these functionalities and can be run on a laptop within 30 minutes.
Availability and Implementation
EGTool is implemented in Python (≥3.8) with dependencies on NumPy, pandas, umap-learn, matplotlib, NetworkX, and scikit-learn. It is installable via pip after cloning the repository. The optional ancestral linkage group detection step (egtool alg) requires R. A Perl script is provided for converting genome annotation files to the input .chrom format. Source code, example data, and tutorials are available at: https://bitbucket.org/viemet/egtool/src/main/