GGRaSP: a R-package for selecting representative genomes using Gaussian mixture models.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Comparative genomic analyses require representative bacterial genome subsets.
• Current methods for subset generation are often supervised and lack consideration for genome attributes.

Purpose of the Study:

• To introduce the Gaussian Genome Representative Selector with Prioritization (GGRaSP) R-package.
• To provide an unsupervised method for generating diverse and prioritized bacterial genome subsets.

Main Methods:

• Utilizes a Gaussian mixture model for unsupervised clustering of genomic relationships.
• Implements a prioritization scheme to retain genomes of interest and minimize genetic variation loss.
• Demonstrated on a dataset of 4600 Escherichia coli genomes, reducing it to 315 representative genomes.

Main Results:

• GGRaSP successfully generates reduced genome subsets while preserving genetic diversity.
• The package allows for prioritization based on user-defined criteria, such as type strains and genome completeness.
• Unsupervised clustering effectively models genomic relationships and aids in threshold selection.

Conclusions:

• GGRaSP offers an efficient and flexible solution for bacterial genome subset selection.
• The R-package facilitates more effective comparative genomic studies by providing representative and prioritized datasets.
• This tool addresses limitations in existing methods for handling large-scale genomic data.