Machine Learning Analysis of Maize Seedling Traits Under Drought Stress

Area of Science:

• Agricultural Science
• Plant Physiology
• Biotechnology

Background:

• Rising greenhouse gas concentrations increase global drought risks, threatening crop productivity.
• Maize (Zea mays L.) is a vital global crop susceptible to drought stress.
• Developing drought-tolerant varieties is crucial for food security.

Purpose of the Study:

• To investigate the effects of drought stress on maize seedling growth and phenotyping.
• To evaluate the predictive performance of machine learning models for identifying drought tolerance traits.
• To provide a foundation for screening drought-tolerant maize and refining breeding strategies.

Main Methods:

• Drought stress was induced in 78 maize hybrids by withholding water for 10 days at the V3 stage.
• Key phenotypic traits assessed included plant height, stem diameter, chlorophyll content, and root number.
• Machine learning algorithms (Random Forest, KNN, XGBoost) were used to analyze traits and predict drought tolerance.

Main Results:

• Plant height, aboveground weight, and chlorophyll content were identified as primary indices for phenotyping drought-stressed maize.
• The XGBoost model achieved high performance in classification (AUC = 0.993) and regression (R² = 0.863).
• XGBoost proved to be the most effective model for predicting maize seedling drought tolerance.

Conclusions:

• Machine learning offers a reliable and feasible approach for screening drought-tolerant maize varieties.
• Phenotypic traits like plant height, weight, and chlorophyll are valuable for assessing drought tolerance.
• This research supports the advancement of precision breeding strategies for enhanced maize resilience.