CLAW-MRM: Comprehensive Lipidomics Automation Workflow for Multiple Reaction Monitoring Using Large Language Models

Lipidomic profiling generates vast datasets, making manual annotation and trend interpretation complex and time-intensive. The structural and chemical diversity of the lipidome further complicates the analysis. While existing tools support targeted lipid identification, they often lack automated workflows and seamless integration with statistical and bioinformatics tools. Here, we introduce the comprehensive lipidomics automated workflow for multiple reaction monitoring (CLAW-MRM), a platform designed to automate lipid annotation, statistical analysis, and data parsing using custom multiple reaction monitoring (MRM) precursor product ion transitions. CLAW-MRM employs trimmed mean of m-value (TMM) normalization to account for lipid load differences, enabling robust cross-sample comparisons. To evaluate CLAW-MRM’s performance, we analyzed lipid profiles in liver tissues of Alzheimer’s disease (AD) mice and age-matched wild-type controls under conditions of constant and variable tissue mass, assessing the impact of normalization strategies on TMM-normalized lipidomic outcomes. Additionally, we isolated and profiled lipid droplets from individual brain regions of 18- to 24-month-old AD male mice and controls, leveraging nearly 1,500 MRM transitions across 11 lipid classes. Enhancing biological relevance, CLAW-MRM integrates LIGER (lipidome gene enrichment reactions), linking lipid expression with gene activation and suppression patterns. Through CLAW-MRM-based LIGER, we identified metabolic pathways enriched in differentially expressed lipids, offering insights into altered lipid metabolism in AD. To improve usability, CLAW-MRM incorporates a natural language interface powered by large language models, enabling artificial intelligence (AI)-driven user interaction for statistical and bioinformatics analyses. By automating lipid structural identification and integrating AI-assisted bioinformatics, CLAW-MRM provides an end-to-end workflow from data acquisition to interpretation, streamlining high-throughput lipidomics.