A GC–MS Data Analysis Platform for Untargeted Metabolomics with Enhanced Coeluting Peak Resolution

Gas chromatography–mass spectrometry (GC–MS)
remains
challenged by the accurate resolution of coeluting peaks and the correction
of retention time shift in large-scale batch analysis. Here, we introduce
AntDAS-CPR, an integrated data analysis platform tailored for untargeted
GC-MS-based metabolomics. The platform incorporates modules for total
ion chromatogram (TIC) peak resolution, retention time shift correction,
component registration, chemometric analysis, and compound identification.
In this work, the TIC peak resolution module was specifically optimized
through the development of a dynamic elimination multivariate curve
resolution–alternating least-squares (DEMCR-ALS) algorithm,
which employs multiple initialization strategies to enhance the resolution
of coeluting peaks and reduce dependence on initial estimates inherent
in conventional methods. The performance of AntDAS-CPR was comprehensively
evaluated using standard mixtures and complex food matrix data sets.
It was compared with state-of-the-art GC–MS data analysis tools,
including AMDIS, ADAP-GC, MS-DIAL, and eRah. Comparative results demonstrated
that AntDAS-CPR consistently outperforms existing methods in both
targeted and untargeted analysis. The platform is freely accessible
at http://www.pmdb.org.cn/antdascpr.