LIPID MAPS Structure Database (LMSD)

Summary

LMSD is the curated structural core of LIPID MAPS. It combines standardized lipid representations and identifiers with names, chemical properties, classifications, source taxonomy, literature provenance, cross-database mappings, and biochemical reactions, supporting both human browsing and programmatic lipidomics workflows [1–3].

Data Model and Search

The original database collected structures from LIPID MAPS laboratories, consortium experiments, computational generation for appropriate classes, and manual curation of public resources and literature. It applied a consistent drawing convention, assigned LM identifiers, and supported classification browsing, text queries by name/formula/mass/class, exact-structure and substructure search, and links to external databases [1].

The distinction between experimentally characterized and computationally generated structures is important. For lipids whose sn positions or other isomeric details were unresolved, early LMSD represented the ambiguity rather than treating every combinatorial structure as experimentally demonstrated [1].

Current Curation and Interoperability

The 2024 update reported more than 48,000 structures, literature provenance for newly curated records, NCBI taxonomy on approximately 18,000 entries, multiple-class browsing for multifunctional lipids, refreshed ChEBI and PubChem mappings, and links to PDBe structures. Around 9,500 glycan-containing molfiles were redrawn in a machine-interpretable planar form while retaining perspective views for display [2].

LMSD also integrates enzyme-catalyzed and non-enzymatic reactions from verified WikiPathways, Rhea, Reactome, and expert curation. Generic class-level reactions can cascade to relevant molecular species while preserving their radyl groups, and the Reactions Explorer exposes the network for navigation [2].

By 2026 LMSD had reached 50,000 curated lipid structures and linked individual entries to experimental data and spectra in resources including Metabolomics Workbench, NP-MRD, and MassBankEU [3].

Identifier Semantics

LM identifiers originally encoded the classification hierarchy. The current policy keeps an identifier stable when a lipid is reclassified, so the identifier is a persistent record key and not a guaranteed current description of class membership [2].

Appropriate Use

• Use LMSD when a biologically relevant structure or well-curated record is required.
• Match search output to the structural resolution of the experiment using the LIPID MAPS shorthand standard.
• For precursor-mass searches with unresolved structure, use species-level results rather than selecting an exact LMSD isomer without supporting fragmentation or orthogonal evidence [2].

Citations

[1] Sud et al. 2007. LMSD: LIPID MAPS structure database. Supports: Original sources, data model, standardized structures, search interfaces, isomer handling, and LM identifiers. Location: Abstract; Database Content and Description; User Interfaces; Summary.

[2] Conroy et al. 2024. LIPID MAPS: update to databases and tools for the lipidomics community. Supports: Structure counts, taxonomy, provenance, multiple classifications, glycan redraw, persistent identifiers, reactions, and bulk-search guidance. Location: Updates to LMSD; Bulk Searches; Programmatic Access.

[3] Cockayne et al. 2026. LIPID MAPS: Powering discovery in lipidomics. Supports: 50,000-structure milestone and links to raw studies, NMR spectra, and MS/MS resources. Location: Main text, page 2.