ReGAIN: a bioinformatics platform for assessing probabilistic co-occurrence between resistance genes in bacterial...

Area of Science:

• Microbiology
• Bioinformatics
• Computational Biology

Background:

• Global rise of multidrug-resistant bacterial pathogens necessitates advanced analytical methods.
• Existing methods lack scalability for inferring co-occurrence and conditional dependencies of resistance determinants.
• Understanding shared genetic context is crucial for combating antimicrobial resistance.

Purpose of the Study:

• To introduce ReGAIN (Resistance Gene Association and Inference Network), an open-source framework.
• To apply Bayesian network structure learning for inferring probabilistic relationships among resistance and other determinants.
• To provide a scalable and reproducible method for population-wide analysis of bacterial resistance networks.

Main Methods:

• Utilized Bayesian network structure learning to model dependencies.
• Analyzed co-occurrence of antibiotic resistance, heavy metal tolerance, stress response, and virulence determinants.
• Applied the framework to ESKAPEE pathogens.

Main Results:

• ReGAIN successfully recapitulated known resistance gene relationships.
• Identified novel candidate patterns indicative of co-selection and shared genetic context.
• Generated conditional probabilities, relative risks, and absolute risk differences for prioritized analysis.

Conclusions:

• ReGAIN offers a powerful tool for scalable, reproducible population-wide analysis of resistance networks.
• The framework supports enhanced surveillance, comparative genomics, and epidemiological studies.
• Facilitates deeper understanding of the genetic basis of antimicrobial resistance.