Machine learning-assisted meta-analysis reveals melatonin crosstalk regulating plant growth plasticity under heavy metal stress.

Heavy metal (HM) pollution is a traditional and ongoing global ecological issue that requires continuous attention. Although studies have focused on how plants cope with HMs stress, the related results ineffectively serve phytoremediation and control HMs toxicity. Melatonin (MT), as a multifunctional signaling molecule, has significant potential in alleviating HMs toxicity in crops. In this meta-analysis, a dataset across different plant families comprising 2476 observations from 140 studies evaluates MT-mediated crosstalk regulating plant growth under HMs stress. Results revealed that exogenous MT significantly improved overall plant performance by 18% (95% CI = 19.5%-14.8%), including plant growth (41.4%), morphology (51.3%) and physiology (44.4%) by stimulating shoot and root biomass 45.5% and 39.5%, respectively. The marked increment in chlorophyll (33.6%) and antioxidant activities (31.8%) was observed by mitigating the adverse effects of oxidative damage (- 23.5%). The elevated nutrient acquisition (24%) and endogenous MT level (59%) were accompanied by 24.2% increase in metabolites and -19.4% reduction in HMs uptake. Random forest machine learning revealed that plant species, MT duration and HMs concentration were the dominant predictors of biomass, whereas HMs type and exposure duration followed by MT level and duration exerted significant influence on plant length and oxidative stress. These effects were significantly associated with MT-mediated genes response and dose or exposure duration, where MT ≤ 100 μM for ≤ 15 days significantly improved results, while excessive level of HMs and prolonged exposure duration diminished MT efficacy. These findings underscore the promising potential of MT-mediated mechanisms in attenuating the injurious effects of HMs on plants.