In-vitro a​nd in-vivo plant growth-promoting assessment of native actinomycetes for enhancing Abelmoschus esculentus L. growth: pattern recognition and machine learning-based recommendations

Food Science of Animal Resources · Springer Nature · 2026-11-09 · DOI: 10.1007/s13205-026-05131-0

Аннотация

Actinomycetes can contribute to plant growth and nutrient acquisition through multiple plant growth-promoting mechanisms. Plants find it challenging to absorb nutrients (N, P, and K) because of the absence of helpful microbes in the soil. In this regard, the current study on the application of actinomycete isolates Sr KAC3 (PP550146), Sb KAC4 (PP177363), Sm KAC6 (PP177364), Ma KAC8 (PP177365), Sc KAC10 (PP177366), Sw KAC11 (PP177374), and Sg KAC12 (PP177389) in different consortia formulations challenged okra plants during Kharif and Rabi seasons of 2023–2024 in order to record promoting plant growth traits (IAA, Phosphate solubilization, potassium solubilization, siderophore and ammonia production), physiological characteristics (chlorophyll content, relative water content), and yield assessment (flowering count, total fruit weight and fruit numbers). The field experiment comprised 3 replicates per treatment (5.92 m²), 11 treatments per site, 10 plants per replicate (1.97 m²), two sites, and two seasons, totally 1,320 plants. The data from two sites at two seasons has been used for machine learning prediction. The Random Forest Regressor model was used to optimize parameters influencing okra growth and anticipate significant treatments to improve precision farming. ANOVA and Kruskal-Wallis tests confirmed robust treatment effects ( p < 0.001), corroborated by Random Forest models achieving R² ≥ 0.998, establishing C 9 as the most effective growth-promoting treatment. Experiments suggested that consortium C 9 ( Streptomyces bangladeshensis (Sb) KAC4, Microbacterium arabinogalactanolyticum (Ma) KAC8, Streptomyces calvus (Sc) KAC10) enhanced plant growth and improved fruit numbers. The Random Forest models describe the relationship between measured traits and yield components within this experiment and were not designed for pre-harvest treatment selection.

Читать у издателя →

Похожие статьи