Research Information System
Euro-Mediterranean Journal for Environmental Integration, vol.11, no.1, 2026 (ESCI, Scopus)
This study was conducted on wild Ajuga iva species, widely used in Tunisian folk medicine for treating various diseases. It was selected as a model plant to introduce an innovative predictive modelling approach that aims to establish a connection between its phenolic composition and related biological activities. For that, six extracts were prepared using solvents of increasing polarity (Hex: Hexane, DCM: Dichloromethane, EtAc: Ethyl acetate, BuOH: Butanol, EtOH: Ethanol, and water: Aq) and evaluated for a range of biological activities. Kaempferol, cynaroside, and gallic, 3,4-di-O-caffeoylquinic, and salvianolic acids were identified by LC–MS analysis, for the first time in the A. iva’s aerial part. Based on chromatographic profiling and response surface methodology (RSM), the established model exhibited robust predictive power (R2 = 0.946) in correlating phenolic content with bioactivity. Variance inflation factor (VIF) analysis allowed selecting salvianolic acid, quercetin, cirsiliol, and cirsilineol as the most impactful contributors to this bioactivity. These findings were further supported by Partial least squares regression (PLSR) to confirm the model’s ability in capturing complex multivariate interactions. Despite this, interpretation of the negative regression coefficients in the generated model, raised questions about underlying compound dynamics, potentially influenced by multicollinearity, synergistic or antagonistic effects, or concentration thresholds. These interactions warrant further investigation through in silico approaches to better understand the mechanistic insights into the real effects.