Rapid screening of bacteriostatic and bactericidal antimicrobial agents against by combining machine learning (artificial intelligence) and UV-VIS spectroscopy.

Antibiotics are compounds that have a particular mode of action upon the microorganism they are targeting. However, discovering and developing new antibiotics is a challenging and timely process. Antibiotic development process can take up to 10-15 years and over $1billion to develop a single new therapeutic product. Rapid screening tools to understand the mode of action of the new antimicrobial agent are considered one of the main bottle necks in the antimicrobial agent development process. Classical approaches require multifarious microbiological methods and they do not capture important biochemical and organism therapeutic-interaction mechanisms. This work aims to provide a rapid antibiotic-antimicrobial biochemical diagnostic tool to reduce the timeframes of therapeutic development, while also generating new biochemical insight into an antimicrobial-therapeutic screening assay in a complex matrix. The work evaluates the effect of antimicrobial action through "traditional" microbiological analysis techniques with a high-throughput rapid analysis method using UV-VIS spectroscopy and chemometrics. Bacteriostatic activity from tetracycline and bactericidal activity from amoxicillin were evaluated on a system using non-resistant O157:H7 by confocal laser scanning microscopy (CLSM), scanning electron microscopy (SEM), and UV-VIS spectroscopy (high-throughput analysis). The data were analysed using principal component analysis (PCA) and support vector machine (SVM) classification. The rapid diagnostic technique could easily identify differences between bacteriostatic and bactericidal mechanisms and was considerably quicker than the "traditional" methods tested.