Towards a Unified Classification of Traditional Bioinputs Using Multi-Omics and Machine Learning Tools
Year: 2025
Type: article
Abstract: The application of traditional bioinputs in agriculture—such as compost, fermented plant extracts, and microbial inoculants—has gained renewed attention for their potential in promoting sustainable and eco-friendly farming. However, their classification and standardization remain fragmented due to regional diversity, variable preparation methods, and lack of consistent molecular characterization. This study proposes a unified classification framework for traditional bioinputs using multi-omics technologies (genomics, transcriptomics, metabolomics, and proteomics) integrated with machine learning tools. We collected and analyzed samples from diverse bioinput types across several agroecological regions. Using high-throughput sequencing and analytical chemistry, we identified distinct microbial, chemical, and genetic signatures. Machine learning models—including Random Forest, Support Vector Machines, and Deep Neural Networks—were trained to classify bioinputs based on their omics profiles. Our results indicate that bioinputs can be reliably grouped into four major categories: microbial-based, plant-extract-based, animal-waste derived, and mixed/fermented consortia. These classifications were validated with over 90% accuracy and demonstrated strong correlations with functional markers of bioactivity. Our findings show that traditional bioinputs, though heterogeneous in preparation, exhibit reproducible molecular patterns that can be standardized. The integration of omics data provides a robust and scalable method for classifying bioinputs, enabling improved quality control, policy formulation, and application guidance. The proposed classification system offers a scientific foundation to support the development and global harmonization of traditional bioinput use in agroecology.
Authors: BHAVANI PAGALLA, SANTHI DASARI, NAVEENA LAVNYA Jeevigunta
Language: English
Subfield: Plant Science
Domain: Life Sciences
Open Access status: hybrid
OpenAlex ID: W4414599376