Detection of microbial coexistence networks in ancient monuments: machine learning and data visualization

The present dissertation investigates the detection and analysis of microbial networks in ancient monuments using machine learning techniques. Initially, urban microbiomes and the environmental factors influencing their composition and functionality are analyzed. The interactions between microorganisms are presented through network analysis, along with modern metagenomics using Next Generation Sequencing (NGS) for studying microbial diversity. Subsequently, species presence in metagenomic data is examined, and the results of microbial analyses are presented using Sankey diagrams and BioLayout networks. In this context, the study expands into the use of machine learning techniques for microbial data analysis. Methods for microorganism classification and decision tree construction are presented and evaluated to optimize the identification and prediction of microorganism presence for selected, abundantly present species, as computed biomarkers. Through the application of these techniques, the potential of machine learning to enhance the understanding of microbial communities and contribute to a more reliable interpretation of metagenomic data is highlighted. Finally, the applications of metagenomics in urban environments are discussed, as well as the prospects arising from the integration of bioinformatics and artificial intelligence in the study of microorganisms.