A0241
Title: Finding groups in microbiome data according to multiple data-views
Authors: Silvia Dallari - Alma mater studiorum- universita di Bologna (Italy) [presenting]
Laura Anderlucci - University of Bologna (Italy)
Angela Montanari - Universita di Bologna (Italy)
Abstract: Microbiota plays a crucial role in human health. Next Generation Sequencing technologies have enabled the exploration of the microbiome without isolation and culturing. However, analyzing and translating microbiome data into meaningful biological insights is still challenging due to the data's compositional nature, high dimensionality, sparseness, and over-dispersion. The gut microbiome can vary from individual to individual, and microbiome communities can be grouped to identify community types linked to environmental or health conditions. Different data features, such as individual profiles, community-based descriptors, or genera interactions within a community, provide different perspectives on microbiome complexity. Combining these perspectives may lead to a more comprehensive understanding of microbiome data. The clustering results of the three data views could be combined via consensus clustering or via Bayesian latent structure models. The proposed multi-view clustering method will be applied to a real dataset on the human gut microbiome.
