Isolation of Health-Benefiting Bacteria from Probiotic Foods

Overview

Probiotic Foods are known to have many health benefits such as bolstering intestinal health, reducing risk to many diseases, and reducing antibiotic resistance, to name a few. Because these foods have so many benefits, I hypothesized that by isolating bacteria from them, these health benefits could be extended to all kinds of different foods. The bacteria of interest were isolated from Great Value Greek Plain Nonfat Yogurt by using MRS medium to encourage the growth of lactic acid bacteria. After several strains had been isolated, they were used to ferment orange juice and carrot juice for about three days. By using HPLC (High-Performance Liquid Chromatography), several differences between the original juices and the fermented juices were found. E. Coli Bacteria were tested against antibiotics as well as salt concentrations after being exposed to one of two treatments: fermented juice and non-fermented juice. Through this experiment, I have been able to observe positive results from three bacterial strains. I was able to observe a decrease in antibiotic resistance to ampicillin, hygromycin, carbenicillin, and kanamycin when I exposed E. Coli bacteria to juices fermented by my isolated bacteria. Constructing a Novel Multidirectional Machine Learning-Based Stochastic Process Model to Simulate the Latitude of the North Atlantic Jet Stream Nicole Ma Sage Hill School, Newport Coast, CA Teacher Todd Haney, Sage Hill School The North Atlantic Jet Stream (NAJS) is greatly influential in the distribution of precipitation and the development of severe weather events, such as cyclones, tropical storms, and hurricanes, throughout North America and Europe. Significant shifts in the latitude of the NAJS can have disastrous impacts on human and environmental well-being. However, current climate models’ simulations of jet stream latitudes are inaccurate due to their reliance on simplified physical parameterizations. Similarly, unidirectional machine learning models lack complexity and do not consider the interconnectedness of the physical drivers of the NAJS. In this study, a set of multidirectional stochastic process models were created to simulate the latitude of the NAJS by applying the machine-learning technique of vector autoregression to times series constructed from a suite of global climate models. Statistical and time series analyses, including cointegration tests and lag selection, were used to determine parameters for the stochastic process models. T o determine the optimal model, the root mean square error (RMSE) of each model was calculated with respect to historical, observational data from the ERA5 global reanalysis. The optimal model was created from MPI-ESM-1-2-HAM data, and its RMSE score demonstrated a significant 36.1% improvement from the RMSE score of the baseline, unidirectional model. These results show that multidirectional models exhibit considerably greater reliability in forecasting changes in jet stream latitudes than traditional unidirectional models. Forecasts from this novel, highly accurate stochastic process model can be applied to forecast future latitudes of the NAJS and facilitate preparation for upcoming catastrophes.