Mandrake: visualizing microbial population structure by embedding millions of genomes into a low-dimensional...

Area of Science:

• Microbiology
• Genomics
• Bioinformatics

Background:

• Microbial population genomics has rapidly expanded, with hundreds of thousands of genomes available for single species.
• Advances in sequencing technology and genomic surveillance initiatives are accelerating data generation.
• Existing methods struggle to explore the vast population structures from diverse genomic data modalities.

Purpose of the Study:

• To present Mandrake, an efficient dimensional reduction method for large-scale population genomics.
• To enable rapid exploration and visualization of microbial population structures.
• To provide accessible tools for analyzing large genomic datasets.

Main Methods:

• Developed Mandrake, an efficient implementation of dimensional reduction tailored for population genomics.
• Applied Mandrake to visualize population structure from multiple sequence alignments, assemblies, and gene content data.
• Utilized several datasets representing major pathogens to demonstrate method's utility.

Main Results:

• Mandrake effectively visualizes population structure from millions of whole genomes.
• Demonstrated the method's usefulness with real-world pathogen datasets.
• The method handles diverse data modalities, including alignments, assemblies, and gene content.

Conclusions:

• Mandrake provides an efficient solution for exploring large-scale microbial population genomic data.
• The tool facilitates understanding of population structures in microbial pathogens.
• Freely available analysis pipeline and interactive web application enhance accessibility.