New approach for rapid detection of bioactive compounds using MIR spectroscopy and machine learning algorithms

Nowadays, microbial infections and resistance to antibiotic drugs are the biggest challenges, which threaten the health of societies. Due to several pharmacological activities associated with Cynara cardunculus, such as hepatoprotective, antioxidative, anticarcinogenic, hypocholesterolemic, antibacterial, anti-HIV, among others, extracts from seeds, leaves, and flowers were tested in Escherichia coli cells. The sensibility of the Mid-infrared (MIR) spectroscopy allowed to perform a detailed analysis of the antimicrobial action of extracts in terms of their biomolecular changes. A comparative model based on several commercial antibiotics such as metronidazole, kanamycin, clarithromycin, chloramphenicol, and ampicillin, was developed. The clustering analysis was performed using unsupervised algorithms such as Principal Component Analysis (PCA), and Kohonen Self-Organizing Maps (SOM). The extracts characterized with antioxidant activity were clustered with antibiotics and presented a promissory antimicrobial activity. According to this preliminary result, it is possible to use the MIR spectroscopy and machine learning algorithm to discover promissory bio compounds characterized by antimicrobial properties, allowing to develop a platform to discover new bioactive molecules, reducing time and costs.