Research
From mass spectra to metabolic function
How can we make metabolomics reproducible and biologically interpretable?
We develop open computational tools for LC–MS data processing, compound annotation and functional analysis. TidyMass provides a modular, traceable workflow; metID automates compound annotation; TidyMass2 extends analysis to metabolite origins and functional modules, including information carried by unannotated features.
AI and networks for multi-omics interpretation
Which molecules act together, and how do their relationships change over time?
We combine biological networks, statistical modelling and AI to turn multi-omics measurements into testable hypotheses. MAPA discovers functional modules and uses literature-grounded language models to interpret them, with applications to aging-related datasets. lagCI examines time-lagged relationships in dense molecular and wearable time series. These inferred relationships guide follow-up experiments rather than establishing causality on their own.
Microbial communities and host metabolism
How do microbial communities across the body relate to human physiology?
Our collaborative studies connect microbial dynamics with host molecular and clinical measurements. Longitudinal profiling across four body sites reveals both stable, individual-specific communities and changes associated with health and disease. Related work examines cross-body-site microbial interactions and their links to the circulating metabolome.
Measuring health as a trajectory
What can repeated molecular measurements reveal beyond a single snapshot?
We study how molecular profiles vary with daily life and environmental exposure. Multi-omics microsampling supports frequent measurements of lifestyle-associated changes, while longitudinal exposome profiling links external exposures with internal molecular responses. Together with wearable data, these approaches motivate personalized monitoring that accounts for each person’s baseline and variability.
Metabolic trajectories of pregnancy
Can metabolic changes help characterize gestational progression?
Our pregnancy studies characterize coordinated metabolic changes in longitudinal blood and urine samples. We develop and evaluate models of gestational age and time to delivery, and examine pathways that change across pregnancy. The aim is to build a molecular understanding of maternal physiology and a foundation for future biomarker validation.
