Mathematical models for predicting protodioscin in tropical forage grasses.

Area of Science:

• Agricultural Science
• Animal Nutrition
• Plant Physiology

Background:

• Protodioscin, a saponin found in tropical grasses like Brachiaria and Panicum, can negatively impact animal production.
• Accurate prediction of protodioscin concentration is crucial for mitigating economic losses in livestock farming.

Purpose of the Study:

• To evaluate the efficacy of mathematical models in predicting protodioscin concentration in tropical grasses.
• To compare the accuracy and precision of Linear, Quadratic, Exponential, and Logarithmic models for protodioscin prediction.

Main Methods:

• Assessed protodioscin concentration in six Panicum and five Brachiaria grass cultivars at various regrowth stages (51-111 days).
• Applied and validated four mathematical models (Linear, Quadratic, Exponential, Logarithmic) to the collected data.
• Utilized pairwise mean squared error and delta Akaike information criterion for model comparison.

Main Results:

• No significant differences in accuracy or precision were found among the evaluated models across all grasses.
• The exponential model demonstrated a strong ability to explain protodioscin concentration variability in relation to plant age for Brachiaria grasses.
• Panicum grasses exhibited a consistent protodioscin concentration, irrespective of plant age.

Conclusions:

• Mathematical models can effectively predict protodioscin concentration in Brachiaria grasses using plant age as a variable.
• The exponential model is recommended for predicting protodioscin concentration in Brachiaria grasses.
• Panicum grasses maintain a stable protodioscin concentration, suggesting different management strategies may be needed.