DIATAGeR: Triacylglycerol Annotation of Data-Independent Acquisition based Lipidomics

Triacylglycerols (TGs) are the most abundant lipids in the human body and the primary source of energy storage. Altered TGs are implicated in metabolic syndrome and detailed structural information may provide novel insights into diseases pathology. TGs are comprised of three fatty acyls with various lengths and double bond composition, complicating structural annotation. DIA-based lipidomics enables a continuous and unbiased acquisition of all TGs, creating the potential for more comprehensive TG analysis. However, DIA generates multiplexed tandem mass spectra (MS2) that pose challenges to identifying TGs. Here, we present DIATAGeR, an R package aimed to automate and improve TG identification to the molecular species level in DIA-based lipidomics. With DIATAGeR, TGs are identified using a TG-centric approach, where each TG in the reference database is considered as an analysis target, searched in DIA spectra, and scored using a logistic regression machine learning algorithm. Additionally, DIATAGeR uses an FDR correction calculated by a target-decoy approach to improve the confidence of TG identification and limit false positives due to interference from unrelated ions. The performance of DIATAGeR was validated in a lipidomic study of liver and plasma samples from mice with metabolic dysfunction-associated steatohepatitis (MASH) and healthy controls. All 9 TG standards were annotated at an FDR < 0.1 in both datasets. When benchmarked against MS-DIAL, TGs identified by DIATAGeR contained 16% and 12% more even-carbon fatty acyls in liver and plasma datasets, respectively. DIATAGeR is a valuable tool for streamlining complex TG annotation in DIA-lipidomics data and freely available at https://github.com/Velenosi-Lab/DIATAGeR.