Assessment of soluble carbohydrate levels in alfalfa and application of artificial intelligence technologies

The aim of this study is to assess the level of neutral detergent soluble carbohydrates of alfalfa (Medicago varia Mart.) when it is cultivated in the Central Non-Chernozem region on soddy-podzolic soil, as well as to review new methods (artificial neural networks) in solving this issue. The content of neutral detergent soluble carbohydrates (NDSC) was determined by the formula: 100 – crude protein (CP) – neutral detergent fiber (NDF) – crude ash (CA) – crude fat (CF). Data on the content of the components of the formula were used to determine levels of NDSC in four alfalfa varieties at the following growth phases: branching, beginning of budding, budding and flowering. Concentrations of NDSC decreased as the plants matured and were (% per dry matter on average for all varieties): at branching – 39.4, at the beginning of budding – 35.5, at budding – 32.7, and at flowering – 26.6. In cooler and more humid conditions of vegetation, the content of CP and NDF was lower, and therefore the level of NDSC was higher and amounted to 36.1% in the budding phase. There is a strong negative correlation between NDSC and the contents of the sum of NDF + CP and NDF alone. The correlation coefficients were 0.96 and 0.90, respectively. Among the components of neutral detergent carbohydrates, the content of non-structural carbohydrates and their proportion were determined. Both indicators decreased with the growth of the plants. Methods for the determination of protein, fiber and crude ash can be based on the use of infrared analyzers. It is important to build appropriate calibration models that are non-linear in nature. Advanced methods for building calibration models include artificial intelligence methods, namely artificial neural networks. Such a methodology can be used, for example, to recommend the optimal period of grass cutting.

1. Homolka P., Koukolova V., Podsedniček M., Hlavačkova A. Nutritive value of red clover and lucerne forages for ruminants estimated by in vitro and in vivo digestibility methods. Czech Jornal of Animal Science. 2012;57(10):454-568. http://doi.org/10.17221/6346-CJAS

2. Rooke John A., Hatfield Ronald D. Biochemistry of Ensiling. URL: http://digitalcommons.unl.edu/usdaarsfacpub/1399 (accessed: August 01, 2024)

3. Hall M.B., Hoover W.H., Jennings J.P. and Miller Webster T.K. A method for partitioning neutral detergent-soluble carbohydrates. Jornal of the Science of Food and Agriculture. 1999;79:2079-2086.

4. Villalba J.J., Ates S and MacAdam J.W. Non-fiber Carbohydrates in Forages and Their nfluence on Beef Production Systems. Frontiers in Sustainable Food Systems. 2021;5:56633875. https://doi.org/10.3389/fsufs.2021.566338

5. Van Soest P.J., Robertson J.B., Lewis B.A. Methods for dietary fiber, neutral detergent fiber, and nonstarch polysaccharides in relation to animal nutrition. The Journal of Dairy Science. 1991;74:3583-3597. https://doi.org/10.3168/jds.S0022-0302(91)78551-2

6. Mary Beth Hall. Working with Non-NDF Carbohydrates with Manure Evaluation and Environmental Considerations. URL: https://www.txanc.org/Proceedings/2002/Non-NDF-Carbohydrates.pdf (accessed: August 01, 2024)

7. Marković J., Babić S., Terzić D., Zornić V. et al. Carbohydrate content of alfalfa harvest at different development stage in the spring growth. 11th International Symposium ‘Modern Trends in Livestock Production’. October 11-13, 2017. Belgrade, Serbia, 2019:706-712.

8. Gu Y., Wu J., Guo Y., Hu S. et al. Grade Classification of Camellia Seed Oil Based on Hyperspectral Imaging Technology. Foods. 2024;13(20):3331. https://doi.org/10.3390/foods13203331

9. Kaewsorn K., Phanomsophon T., Maichoon P., Pokhrel D.R. et al. Modeling Textural Properties of Cooked Germinated Brown Rice Using the near-Infrared Spectra of Whole Grain. Foods. 2023;12(24):4516. https://doi.org/10.3390/foods12244516