Investigating Potential Anti-Bacterial Natural Products Based on Ayurvedic Formulae Using Supervised Network Analysis and Machine Learning Approaches.

Abstract

Objectives: This study implements a multi-dimensional methodology to systematically identify potential natural antibiotics derived from the medicinal plants utilized in Ayurvedic practices. Methods: Two primary analytical techniques are employed to explore the antibiotic potential of the medicinal plants. The initial approach utilizes a supervised network analysis, which involves the application of distance measurement algorithms to scrutinize the interconnectivity and relational patterns within the network derived from Ayurvedic formulae. Results: 39 candidate plants with potential natural antibiotic properties were identified. The second approach leverages advanced machine learning techniques, particularly focusing on feature extraction and pattern recognition. This approach yielded a list of 32 plants exhibiting characteristics indicative of natural antibiotics. A key finding of this research is the identification of 17 plants that were consistently recognized by both analytical methods. These plants are well-documented in existing literature for their antibacterial properties, either directly or through their bioactive compounds, which suggests a strong validation of the study's methodology. By synergizing network analysis with machine learning, this study provides a rigorous and multi-faceted examination of Ayurvedic medicinal plants, significantly contributing to the identification of natural antibiotic candidates. Conclusions: This research not only reinforces the potential of traditional medicine as a source for new therapeutics but also demonstrates the effectiveness of combining classical and contemporary analytical techniques to explore complex biological datasets.