LAUSR

BACKGROUND The soil microorganisms can form complex and varied communities which interact with each other in many different ways depend to environment conditions. These microbial diversities are accompanied by different metabolic paths and adaptability reflected even in extreme environments. In the recent decades, biodiversity of microbes in extreme environments has been in scientific focus because the specifically adapted bacteria can improve bioremediation processes in industrial and agriculture application. Instead of time-consuming process of identification of new bacterial strains from habitats rich in microbiome, artificial neural network was proposed as a mapping model for resolving the problem of prediction of microbial behaviour. RESULT In this article, the occurrence and diversity of alkaliphilic sporogenic bacteria in alkaline soils were investigated. For this purpose, soil samples were collected from different locations: leached soil from the Danube river, cement factory wastewater accumulation, deposit of limestone near the Bešenovo lake and the Beli Majdan cave in the Fruška gora mountain. According to the obtained results, two empirical models were developed that gave a good fit to experimental data and was able to predict the pH and temperature growth profile of the natural isolates successfully. The artificial neural network models showed a reasonably good predictive capability (overall R2 for temperature growth profile was 0.727, while the overall R2 for pH growth profile was 0.906). CONCLUSION The developed mathematical models provided adequate precision for practical study in the microbiology laboratory and scale-up processes for wide range of laboratory and industrial applications, where specifically-adapted microbial community is needed. This article is protected by copyright. All rights reserved.