CALDERA: finding all significant de Bruijn subgraphs for bacterial GWAS

Area of Science:

• Genomics
• Computational Biology
• Microbial Genetics

Background:

• Genome-wide association studies (GWAS) are crucial for identifying genetic variants linked to bacterial traits such as drug resistance and hypervirulence.
• Current k-mer based GWAS methods in bacteria can suffer from diluted effects due to gene variations across strains.
• Existing approaches often require a reference genome, limiting the analysis of accessory genes.

Purpose of the Study:

• To address the limitations of k-mer based GWAS in bacteria by developing a novel method to capture polymorphic genes as single entities.
• To improve the power and interpretability of bacterial GWAS.
• To enhance computational efficiency in analyzing large genomic datasets.

Main Methods:

• Developed a novel approach utilizing closed connected subgraphs (CCSs) from de Bruijn graphs of genomic k-mers.
• Implemented a testable hypothesis framework with a new enumeration scheme for CCSs to manage computational complexity and multiple testing.
• Integrated the method with existing visualization tools for enhanced interpretation.

Main Results:

• The proposed method effectively captures polymorphic genes, overcoming the issue of diluted effects seen in traditional k-mer GWAS.
• Achieved significant improvements in both statistical power and interpretability of GWAS results.
• Demonstrated drastic improvements in computational efficiency through an optimized CCS enumeration scheme.

Conclusions:

• The novel CCS-based GWAS method offers a more robust and efficient approach for analyzing bacterial genomes.
• This method enhances the discovery of genetic determinants for important bacterial phenotypes.
• The provided implementation and code facilitate reproducibility and further research in bacterial genomics.