Analyzing Microbiome Data Using Machine Learning

Microbial communities are integral components of earths ecosystems. Investigating their temporal and spatial distribution patterns provides insights into the dynamics of the entire ecosystem. 

Analyzing amplicon- and metagenome-based microbiome data sets can uncover these patterns. However, these high-dimensional datasets are usually compositional, sparse, and characterized by non-linear relationships, complicating their analysis through conventional statistical methods. 

The application of various machine learning techniques, for example natural language processing algorithms, can circumvent these challenges. These methods facilitate the extraction of essential ecological information and the identification of interpretable patterns, thereby enhancing the understanding of microbial communities and the respective ecosystems. 

Key Publications

Kujat A.S., Glackin C., Hassenrück C.,, Vogel L., Zschaubitz E., Lüdtke S., Labrenz M., Sperlea T. (2024): Detecting interpretable patterns in estuarine and coastal microbial communities using computational linguistics and machine learning methods. The 7th joint conference Joint Microbiology & Infection Conference 2024 of DGHM &VAAM, Würzburg, June 2-5, 2024 (Oral presentation) 

Contact

Anna Kujat

anna.kujat@uni-rostock.de