Abstract: Human microbiome research has expanded rapidly following the advances in metagenomic sequencing and other high-through profiling technologies. The specific properties of microbiome profiling data, such as compositionality, hierarchical phylogenies, and high levels of temporal, spatial, and individual variation set fresh challenges for statistical modeling. This talk provides an overview of contemporary modeling challenges and recent applications of probabilistic machine learning in microbiome data science.

Bio: Leo Lahti is Academy Research Fellow and associate professor in data science in University of Turku, Finland. His research aims to bridge the gap between theory and practice of algorithmic data analysis. Lahti obtained D.Sc. in Aalto University, Finland in 2010, followed by an extended research period in The Netherlands and Belgium focusing on computational analyses of the human microbiome variation in large population studies. Lahti has published widely used open research software, coordinates a related COST action, and organized various international data science training events.

Speakers:  Leo Lahti

Affiliation: Department of Future Technologies, University of Turku

Place of Seminar:  Zoom (Now available on YouTube)