LipidIN

LipidIN enables flash platform-independent annotation and reverse lipidomics through a large lipid fragmentation library and fast querying module for improved annotation accuracy and depth.

Key Features:

• Comprehensive Fragmentation Library: A 168.5-million lipid fragmentation hierarchical library encompassing all potential chain compositions and carbon-carbon double bond locations.
• Expeditious Querying Module: Achieves speeds exceeding one hundred billion queries per second across all mass spectral libraries.
• Lipid Categories Intelligence Model: Developed using three relative retention time rules to reduce false positive annotations and predict unannotated lipids with a 5.7% estimated false discovery rate, covering 8923 lipids across various species.
• Wide-spectrum Modeling Yield Network: Regenerates lipid fragment fingerprints to improve accuracy and coverage with a 20% estimated recall boosting.

Scientific Applications:

• Lipid Annotation: Improves annotation accuracy, coverage, speed and depth of lipid profiles across multiple mass spectrometry platforms.
• Biomarker Discovery: Demonstrated utility in clinical cohorts for identifying lipid biomarkers.
• Unannotated Lipid Prediction: Predicts unannotated lipids with controlled false discovery rate across various species.

Methodology:

LipidIN employs a hierarchical lipid fragmentation library, an expeditious querying module for mass spectral library searches, a lipid categories intelligence model based on three relative retention time rules, and a Wide-spectrum Modeling Yield network for regenerating lipid fragment fingerprints.