Estudo da interação entre híbridos de milho e o fungo endofítico Trichoderma harzianum

Abstract

This study explores sustainable corn production in the context of the challenges posed by climate change and the increasing global food demand. The aim of this study was to evaluate the effect of inoculation with the endophytic fungus Trichoderma harzianum on commercial corn cultivars. Field trials were conducted in the experimental area of UNESP, Ilha Solteira, Brazil, in a humid tropical climate. Single-cross hybrids of corn commonly found in Brazil's major producing regions (DKB360, DKB255, and 2B810PW) were selected; the T. harzianum strain (ESALQ 1306) was used for inoculation at a dose of 2 mL per kg-1 of seeds. Two harvests were evaluated (2020/2021 and 2021/2022), with the last harvest experiencing severe drought. The methodologies used in this study included morphological analyses, chemical composition, and yield in response to inoculation. The collected data were used to develop a machine learning model based on the 'Random Forest' algorithm to classify the evaluated samples as 'inoculated' or 'non-inoculated' based on models trained with the data, using the Python programming language. Kinematic analysis of plant growth in response to inoculation was also performed to monitor cell division and expansion processes to verify the potential capacity of the strain to promote plant growth. The results show that inoculation with T. harzianum strain (ESALQ 1306) increased productivity and nutritional content for all evaluated hybrids, especially in the presence of water stress; however, these effects vary according to the hybrid, with the DKB360 hybrid showing the greatest responsiveness to inoculation, presenting an average increase in grain production of approximately 60%, while the smallest percentage increase in productivity was observed for the 2B810 hybrid, with only a 7% increase in the first year of the study. These data demonstrate that the choice of corn hybrid is crucial for achieving higher productive efficiency in relation to inoculation. The data obtained and analyzed through the “Random Forest” machine learning algorithm enabled the construction of a model capable of classifying with high precision samples belonging to the “inoculated” or “non-inoculated” classes. Thus, this innovative methodology allows for corrective intervention through 'foliar re-inoculation' in case of observed failures, thereby reducing yield losses by early detection and correction of deviations from the expected patterns in inoculated samples. Kinematic growth analysis showed that inoculation promoted an increase in leaf elongation rates, especially in the presence of water stress, with a significant increase in the rate of cell production and in the size of the meristematic zone of the leaves. These findings demonstrate that inoculation of plants with the T. harzianum strain (ESALQ 1306) is indeed capable of promoting plant growth even under water stress conditions; however, the effectiveness of this effect varies according to the genetic makeup of the plant.