Biostimulant effect of seaweed extracts on micropropagation of important ground macrophyte Micranthemum tweediei and data-driven insights using machine learning and response surface regression
摘要
Commercially viable in vitro cultivation of ornamental and aquatic plants depends heavily on the proper use of bio-stimulants that encourage healthy and uniform growth. This study explores the role of two seaweed extracts, Jania rubens and Cystoseira barbata, in promoting in vitro micropropagation of Micranthemum tweediei (Monte Carlo), a widely used aquatic plant. Culture medium was enriched with different concentrations (2.5–20.0%) of J. rubens and C. barbata extracts with 1–4 g L−1 Murashige and Skoog (MS) nutrient media. Considering all results, J. rubens extract was more effective than C. barbata in terms of micropropagation and chlorophyll content of M. tweediei. Addition of 5% J. rubens extract to 2 g L−1 MS medium (approx. half strength) resulted in a 47% increase in clump diameter (5.18 cm), a 173% increase in rooting rate (92.4%), and a 108% increase in shoot clump fresh weight (3.6 g), and 84%, 115%, and 87% increases in chlorophyll a, b, and total chlorophyll levels, respectively, compared to the control 2 g L−1 MS medium without extract. Micropropagation parameters resulting from 4 g L−1 MS (approx. full strength) supplemented with J. rubens extract were lower than those obtained from 2 g L−1 MS. Unlike J. rubens, C. barbata showed the highest micropropagation when 10% extract was added to 4 g L−1 MS nutrient medium, resulting in a clump diameter of 5.11 cm, 42.8% rooting, a fresh clump weight of 7.37 g, and a dry clump weight of 0.59 g. Propagated in vitro plants were easily acclimated to external conditions in aquariums. Results of response surface regression analysis (RSRA) confirmed the results, and MS concentration influenced the plant biomass and chlorophyll contents, whereas seaweed type and concentration regulated the clump diameter and rooting. Application of machine learning models validated and predicted the outcomes precisely. Multi-layer perceptron (MLP) model exhibited superior predictive accuracy for morphological traits, while more predictive accuracy of chlorophyll parameters was attributed to Random Forest (RF) model. This study has shown for the first time that aquatic plants can be micropropagated more successfully and at lower cost by adding seaweed extracts to in vitro nutrient media without the need for growth regulators.