LORA

LORA performs lipid over-representation analysis by translating lipid names with the Java-based Goslin framework into standardized hierarchical nomenclature and identifying structural features enriched in lipidomic datasets.

Key Features:

• Lipid name standardization: Translates lipid names using the Java-based Goslin framework into standardized nomenclatures capturing hierarchical information.
• Grammar selection for normalization: Supports selection of specific grammars for consistent lipid name normalization across datasets.
• Hierarchical lipid representation: Captures headgroups, acyl chains, and modifications up to a complete structural level.
• Comparative dataset analysis: Processes experimental query datasets alongside reference datasets for comparative analysis.
• Over-representation analysis: Identifies lipids significantly enriched in the query dataset relative to the reference set based on user-defined criteria and shared structural features.
• Very Important Lipids (VILs): Defines VILs as lipids present in the most over-represented terms with the highest number of enriched shared structural features (e.g., enrichment of glycerophospholipids containing acyl 20:4 at the sn-2 position; example VILs include PC 18:2/20:4;O and PE 16:0/20:4(5,8,10,14);OH).
• Visualization and reporting: Produces graphical visualizations based on the lipidome hierarchy and compiles analysis results into a report.

Scientific Applications:

• Epilipidomics mining: Enables systematic identification of enriched lipid structural features in epilipidomics datasets.
• Molecular-level interpretation: Facilitates interpretation of lipidomic changes at the level of headgroups, acyl chains, and modifications.
• Metabolomics and systems biology: Supports comparative lipidomics analyses used in metabolomics and systems biology studies.
• Lipid-related disease research: Aids identification of lipid species and structural patterns potentially relevant to disease studies.

Methodology:

Users provide experimental query and reference datasets; select grammars for lipid name normalization; lipid names are translated via the Java-based Goslin framework into standardized hierarchical nomenclature capturing headgroups, acyl chains, and modifications; the tool performs lipid over-representation analysis based on user-defined criteria to identify enriched lipids and shared structural features, defines Very Important Lipids (VILs) from the most over-represented terms, and generates lipidome-hierarchy visualizations and a compiled report.