Optimization and validation of cuminaldehyde in combination with antibiotics against Pseudomonas aeruginosa: An Artificial intelligence driven approach

Pseudomonas aeruginosa is a common cause of infections and is known for its ability to develop resistance to multiple antibiotics. While standard antimicrobial therapies are often effective against planktonic cells, biofilm-associated infections remain difficult to manage. Combination approaches have been explored as a possible way to improve treatment outcomes compared to single-agent therapies. In this study, the antibacterial activity of cuminaldehyde (CA) in combination with ciprofloxacin (CF) and tetracycline (TE) was examined against P. aeruginosa . To investigate potential combinatorial effects, a range of machine learning models including multiple linear regressions, polynomial regression, support vector regression, random forest, and artificial neural networks were utilized. These approaches were integrated with genetic algorithm based optimization to systematically identify optimal dose combinations for effective biofilm inhibition. Among the models tested, the random forest approach showed comparatively better predictive performance (Train R² = 0.98; Test R² = 0.81). Experimental validation of the predicted combination (CA 45 µg/mL, CF 0.01 µg/mL, TE 0.004 µg/mL) resulted in approximately 82% biofilm inhibition under in vitro conditions. These results suggest that machine learning-based approaches may be useful for identifying potentially effective combinatorial dosing strategies for P. aeruginosa biofilm inhibition, although further studies are needed to assess their broader applicability.