← All publications
Comparison & QC
LipidSpace — exploration, reanalysis and QC of large studies
The idea
Comparing thousands of distinct lipidomes is hard. LipidSpace builds a structural space from the lipids themselves, then uses it to measure how similar two lipidomes are — turning a pile of tables into an interpretable map of an experiment.
Key concepts
Graph-based comparison of lipid structures yields a structural space model.
Lipidome similarity is computed from distances within that space.
Machine-learning selects lipid subsets that explain study variables such as health condition.
What we found
Enables rapid (re)analysis and quality control across large-scale studies.
Identifies the lipids that shape each lipidome.
Reads Goslin-normalized names and imports mzTab-M; GUI with tutorials and PDF export.