skDER and CiDDER: two scalable approaches for microbial genome dereplication

Area of Science:

• Genomics
• Bioinformatics
• Microbial Evolution

Background:

• Vast numbers of microbial genomes are available, posing computational challenges for comparative genomics.
• Over-representation of certain lineages in genomic datasets can bias evolutionary analyses.
• Efficient selection of representative genomes is crucial for accurate pangenome investigations.

Purpose of the Study:

• To develop and present skDER and CiDDER, novel tools for genomic dereplication.
• To enable efficient selection of representative microbial genomes for downstream analyses.
• To reduce computational burden and mitigate biases in comparative genomic studies.

Main Methods:

• skDER utilizes nucleotide sequence comparisons and average nucleotide identity (ANI) inference for dereplication.
• CiDDER employs saturation assessment of distinct protein-coding genes for genome selection.
• Tools incorporate features for parameter testing, automated genome downloading, and filtering of plasmids/phages.

Main Results:

• skDER demonstrates efficiency comparable to existing ANI-based tools, maintaining pangenome coverage and adhering to user-defined ANI and aligned fraction (AF) cutoffs.
• CiDDER offers a viable alternative to ANI-based methods, facilitating direct optimization for broad pangenome representation.
• Benchmarking confirms skDER's performance and usability for large-scale microbial genome analysis.

Conclusions:

• skDER and CiDDER effectively address the need for genomic dereplication in microbial research.
• These tools enhance the efficiency and accuracy of comparative genomic studies.
• The developed methods support broader evolutionary and pangenome investigations by providing curated genome subsets.