Artificial Intelligence Assisted Detection of Yeast Fungi with Gelatin-based Graphite Hydrogel Coated Impedimetric Odor Biosensor

Abstract

Infections caused by yeast fungi are common in humans and can cause serious health problems and even death if not treated appropriately and early. In conventional detection methods, yeast fungus culture is performed to determine the strain of the fungus that is the source of infection. However, culture and microbiological strain identification under aerobic conditions takes between 24 hours and 72 hours depending on the sample source, and this long period of time prevents rapid initiation of appropriate treatment. Fast, easy, low-cost and highly accurate identification of the source strain of yeast fungal infection will prevent health problems and deaths caused by yeast fungal infections and reduce unnecessary treatment costs. In this study, which is based on the mechanism of distinguishing odors with different electrical signals generated by the binding of different odor molecules to receptors in different ways, a gelatin-based hydrogel coating was used to simulate the binding pockets provided by various amino acids in odor receptors. A gelatin hydrogel coating mixture was obtained with glycerol, which is present in the hydrogel and holds the coating structure together, and graphite powder, which enhances the electrical properties of the coating. An impedimetric odor sensor was obtained by coating copper wires with gelatin-based hydrogel coating. The interaction of the odor molecules produced in the culture of Candida albicans, Candida glabrata, Candida tropicalis yeast fungi with the impedimetric odor sensor was ensured in a closed environment in a petri dish and the impedance changes in the sensor were monitored with an impedance analyzer. Random Forest machine learning algorithm was used to classify these yeast strains with 94% accuracy. With further studies, if this system is further optimized, it will be possible to diagnose yeast fungal infections in the clinic quickly, easily and with high accuracy.