In a controlled plant experiment, inoculation with the arbuscular mycorrhizal fungus Funneliformis mosseae combined with 250 ppm titanium dioxide nanoparticles produced the strongest measured growth outcomes among drought-stressed Melissa officinalis , or lemon balm. Compared with the control, the combined treatment increased total fresh weight by 21.7%, total chlorophyll by 10.0%, shoot length by 30.2% and protein content by 14.7%. It also reduced malondialdehyde by 39.1%.
The researchers suggested that fungal symbiosis and TiO₂ nanoparticles acted synergistically to mitigate oxidative stress and sustain growth under drought. The study tested drought-stressed plants exposed to TiO₂ nanoparticles at concentrations of 0, 250 or 500 ppm, with or without F. mosseae inoculation, and evaluated the treatments alongside well-watered plants. Because the dataset was limited and shoot fresh-weight measurements were unevenly distributed, they also applied SMOGN augmentation.
The experimental scope was specific to lemon balm, the tested fungal association, three TiO₂ concentrations and the imposed watering conditions. Antioxidant enzyme activity and photosynthetic measurements showed that treatment responses varied with both water availability and fungal status. Among drought-stressed plants, F. mosseae plus 250 ppm TiO₂ nanoparticles was the most effective treatment overall for growth and measured stress indicators.
Under well-watered conditions, plants without the fungus but exposed to 250 ppm TiO₂ nanoparticles recorded the highest ascorbate peroxidase activity, at approximately twice the control level. Across the broader comparison with the control, peroxidase activity increased by 53.0–210.0%, while superoxide dismutase activity rose by 7.6–159.4%. Chlorophyll b increased by as much as 37.9%, and the chlorophyll b-to-a ratio was 14–46% higher across the reported treatment comparisons.
These results indicate that the effect of TiO₂ nanoparticles was not uniform across every combination of water availability and inoculation. Bayesian Ridge regression achieved a reported R² of 0.9194 for predicting shoot fresh weight. The model identified chlorophyll a and chlorophyll b as its most influential predictors, indicating that the algorithm captured strong associations between photosynthetic pigment measurements and the predicted growth outcome.
The wider use of 20 algorithms was intended to reveal relationships in a multifactorial dataset that conventional analytical approaches might overlook. The study therefore presented machine learning as an additional way to model plant physiological responses, while its experimental comparisons remained the source of the reported treatment outcomes. The findings support a measured association between fungal inoculation, TiO₂ nanoparticle exposure and improved plant performance under the tested drought conditions, but they do not establish how the combined treatment would perform in other species, at other doses or beyond the experimental setting.
The limited sample size and imbalance in shoot fresh-weight data also required augmentation before model development, making independent replication important for evaluating the reported predictive performance. Within those limits, the experiment identified 250 ppm TiO₂ nanoparticles alongside F. mosseae as the leading treatment for drought-stressed lemon balm, while Bayesian Ridge provided a high-fidelity prediction of shoot fresh weight from chlorophyll measurements.
Subject of Research: Machine-learning analysis of lemon balm responses to drought, mycorrhizal inoculation and TiO₂ nanoparticles
Article Title: Funneliformismosseae symbiosis and TiO2 nanoparticles under drought in Melissa officinalis L.: a machine learning modeling approach
Article References: Dayani, S., Hosseini, R., & Mazaheri-Tirani, M. (2026). Funneliformismosseae symbiosis and TiO2 nanoparticles under drought in Melissa officinalis L.: a machine learning modeling approach. BMC Plant Biology. https://doi.org/10.1186/s12870-026-09847-2
Image Credits: AI Generated
DOI: 10.1186/s12870-026-09847-2
Keywords: Lemon balm, Plant drought stress, Arbuscular mycorrhizal fungi, TiO₂ nanoparticles, Plant physiology, Antioxidant enzymes, Machine learning, Bayesian Ridge regression, Chlorophyll, Oxidative stress, Funneliformismosseae, symbiosis
Cite Scienmag News
Roger Howard. (October 26, 2026). Machine learning assesses lemon balm responses to drought, fungi and TiO₂ nanoparticles. Scienmag. https://scienmag.com/machine-learning-assesses-lemon-balm-responses-to-drought-fungi-and-tio%e2%82%82-nanoparticles/
Roger Howard. "Machine learning assesses lemon balm responses to drought, fungi and TiO₂ nanoparticles." Scienmag, 26 October 2026, https://scienmag.com/machine-learning-assesses-lemon-balm-responses-to-drought-fungi-and-tio%e2%82%82-nanoparticles/. Accessed 26 October 2026.
Roger Howard. "Machine learning assesses lemon balm responses to drought, fungi and TiO₂ nanoparticles." Scienmag. October 26, 2026. https://scienmag.com/machine-learning-assesses-lemon-balm-responses-to-drought-fungi-and-tio%e2%82%82-nanoparticles/
