Evolutionary Computation techniques for the analysis of Antibiotic Resistance in bacterial infectants

Studying bacteria and viruses is crucial for improving cures and preventing large-scale epidemics or pandemics. Antimicrobial resistance (AMR) refers to the ability of microorganisms to withstand the effects of antimicrobial treatments and poses a significant public health threat, so understanding its prevalence, mechanisms, and spread is a global priority. Research in this sector brought significant advancements in technology, which have resulted in remarkable cost reduction and accelerated processing speeds. Despite these improvements, the acquisition of such technologies can still pose a considerable financial burden in low-resource settings. There remains ample room for further enhancements, particularly through the integration of machine learning and computational intelligence strategies to automate processes. In the past, algorithmic solutions to bacterial resistance were not feasible due to the complexity of microbial genomes, limited data availability, and lack of computational power. Although classical algorithms are still impractical for such complex problems, the increased data and computational power have enabled initial attempts. On the other hand, no attempts involving evolutionary computation have been attempted, despite it offering several advantages compared to machine learning approaches. Firstly, it excels at exploring vast solution spaces, making it effective in complex problems with high-dimensional search spaces. It can handle non-differentiable and discrete problems, making it suitable for tasks involving categorical variables or combinatorial optimization. Additionally, evolutionary algorithms are robust to noisy and incomplete data, producing meaningful results even when data quality is imperfect. They are designed for global optimization, avoiding getting trapped in local optima. Evolutionary algorithms are adaptable to dynamic environments, adjusting to changing conditions. Moreover, they can be parallelized for faster convergence and scalability, making them suitable for complex real-world scenarios. For these reasons the primary objective of this research is to utilize Evolutionary Computation to address the complex issue of bacterial resistance. Specifically, the algorithm aims to identify distinct markers that are highly prevalent in resistant variants but largely absent in susceptible ones. These markers have the potential to be correlated with specific types of resistance, offering valuable insights into the underlying mechanisms of bacterial resistance. Moreover, this research endeavors to contribute to the development of targeted interventions by analyzing and uncovering the significance of these markers. A critical factor influencing the final outcome is the ability to identify a wide range of well-established markers associated with a particular bacterium and a specific antibiotic. This focus on discovering numerous known markers is vital for improving the accuracy and reliability of future research in this important area.