Chitosan nanoparticles induced germination and auxin-related gene regulations in cotton (Gossypium hirsutum L.) using random forest-based multi-output regression analysis

Chitosan nanoparticles (CNPs) are emerging nanobiostimulant known to enhance germination, stress tolerance, and hormonal regulation in plants, yet their role in auxin-related gene regulation in cotton remains unexplored. In this study, cotton seeds were exposed to varying concentrations of CNP (0–2.5 mg/L) under in vitro conditions. Seed germination was counted daily for a continuous 21-day period, followed by analysis using the GerminaR statistical package to generate various germination metrics. The germination metrics were analyzed by Random Forest-based multi-output regression analysis. Auxin-related gene expression profiling of ARF, AUX, and UGT was performed at two different concentrations and two time points, using qPCR normalized to His3. Taguchi design (TD) was employed for analyzing gene expressions to quantify factor significance and noise robustness through the S/N ratio. Results revealed maximum germination percentage (GRP) at 1.0 mg/L CNP. Concentrations above 1.5 mg/L CNP negatively impacted germination kinetics with increased germination speed (GSP) and coefficient of variation in germination (CVG). The multi-output regression analysis revealed the highest predictive performance for variance in germination time (VGT) with an R 2 score of 0.785, followed by 0.746 for the CVG. Gene expression profiling revealed significant upregulation of ARF and AUX genes with increasing concentration and time. Whereas UGT regulation was dependent on both time and concentration, with variable impact. The TD analysis identified concentration as most significant for ARF regulation, while time was significant for AUX and UGT expressions, confirming the role of CNPs as a bio-stimulant. • First integrated study combining CNPs, Random Forest regression, and Taguchi design to analyze cotton germination and auxin genes. • Machine learning showed strong predictive performance (R² up to 0.785) in the simultaneous modeling of seven germination traits. • Auxin-related genes (ARF, AUX, UGT) showed clear concentration- and time-dependent expression patterns. • Taguchi analysis showed concentration dominated ARF regulation, while time mainly influenced AUX and UGT.