LipidIN 2.0: an end-to-end intelligent analysis platform for 4D lipidomics and glycolipidomics

Accurate, comprehensive, and high-throughput lipid annotation and quantitation remains a long-standing challenge due to the complexity of lipid structures and the limitations of existing analytical pipelines. To address these challenges, we introduce LipidIN 2.0, an end-to-end intelligent analysis platform that integrates annotation, rapid quantitative analysis, and statistical modeling in a platform-independent software solution. LipidIN 2.0 incorporates a 972.1-million dynamic hierarchical lipid fragmentation library, covering all potential chain compositions, including canonical lipids, glycolipids, and amidated modified lipids. The optimized expeditious querying module delivers unprecedented computational performance, enabling speeds exceeding one trillion queries per second across large-scale mass spectral databases. To reduce false annotations, LipidIN 2.0 employs a dedicated module that integrates precise predictions of retention time (Rt) and collision cross-section (CCS), achieving a median relative error (MedRE) < 0.3%, which results in an estimated 5.7% false discovery rate while enabling the annotation of 8,923 lipids across multiple species. Additionally, the Wide-Spectrum Modeling Yield network regenerates lipid fragment fingerprints to enhance both recall and coverage, yielding an estimated 20% improvement in annotation recovery. The platform also features a next-generation quantification module, supporting rapid, accurate, and reproducible lipid quantification. We further demonstrate the broad utility of LipidIN 2.0 across a wide range of lipidomics applications, including large-scale lipid annotation, high-confidence biomarker discovery, and translational studies in clinical cohorts. Together, LipidIN 2.0 represents a significant advance toward reliable, efficient, and comprehensive lipidomics analysis.