Predicting population dynamics of antimicrobial resistance using mechanistic modeling and machine learning

Area of Science:

• Microbiology
• Computational Biology
• Public Health

Background:

• Antimicrobial resistance (AMR) infections are a growing global public health concern.
• The development of new antibiotics is hindered by rapid resistance emergence, economic challenges, and regulatory issues.
• Current strategies include antibiotic stewardship and drug repurposing.

Purpose of the Study:

• To review the application of quantitative modeling in understanding and predicting AMR population dynamics.
• To explore the potential of mechanistic and machine learning (ML) models in combating the AMR crisis.
• To discuss the translational value of computational models for guiding treatment design.

Main Methods:

• Review of current literature on mechanistic and machine learning (ML) models for AMR.
• Analysis of challenges in mechanistic model construction.
• Exploration of ML's capacity to address limitations in AMR modeling.

Main Results:

• Quantitative modeling can identify key mechanisms and consequences of AMR development.
• Models can predict resistance persistence and inform treatment strategies.
• ML models show promise in overcoming limitations of traditional mechanistic models.

Conclusions:

• Computational models are valuable tools for understanding and predicting AMR.
• ML offers advanced capabilities for AMR dynamics prediction.
• Further development and application of these models can aid in managing the AMR crisis.