Artificially intelligent soil quality and health indices for 'next generation' food production systems

0546792 - MBÚ 2022 RIV GB eng J - Článek v odborném periodiku
Gomes Zuppa de Andrade, V. H. - Redmile-Gordon, M. - Groenner Barbosa, B. H. - Andreote, F. D. - Wurdig Roesch, L. F. - Pylro, Victor Satler
Artificially intelligent soil quality and health indices for 'next generation' food production systems.
Trends in Food Science & Technology. Roč. 107, JAN 2021 (2021), s. 195-200. ISSN 0924-2244. E-ISSN 1879-3053
Institucionální podpora: RVO:61388971
Klíčová slova: Artificial intelligence * Microbiome * Soil quality * Soil health index
Obor OECD: Microbiology
Impakt faktor: 16.002, rok: 2021
Způsob publikování: Omezený přístup
https://www.sciencedirect.com/science/article/pii/S0924224420306415?via%3Dihub

Currently, the lack of a universal soil quality index (SQI) limits adoption of such an approach and may hinder improvements to crop productivity and environmental sustainability. Some SQIs rely only on physicochemical characteristics, which are slow to change and thus have low sensitivity in predicting soil degradation in an appropriate timescale. In contrast, microorganisms respond quickly to changes in land-use and/or management. Furthermore, microbes generate the enzymes and biophysical structures required for many soil functions which thus drive 'fertility', 'health', and 'quality'. Therefore, understanding of community-driven transformations should enable prediction of the trajectories of soil quality in response to management. However, the multitude of varied consequences and feedback loops which emerge dependent on site-specific factors are beyond the capability of models that currently exist. Enormous amounts of soil (meta)genomic data has been generated in the last decade. In parallel, advances in Artificial Intelligence (AI) have revolutionized our capacity to create predictive models in several areas, such as helping plant breeders searching for specific beneficial traits, and informing crop-management by predicting changes in the weather. As soil microbiologists and bioinformaticians, we contend that creating a universal, robust and dynamic Artificially Intelligent Soil Quality Index (AISQI) implies taking advantage of machine learning algorithms and soil microbiome data together with conventional physicochemical parameters and productivity data. This index must be flexible enough to encompass regional peculiarities but allow for comparative studies. Refining different models within the same index might improve its accuracy helping make real-time predictions. The establishment of a collaborative effort is fundamental to creating this index with maximum utility in enhancing agricultural management practices and ecosystem sustainability.
Trvalý link: http://hdl.handle.net/11104/0323173