Herbal Component Correlation and Matrix-based Resolution in Comprehensive two-dimensional Gas Chromatography - Mass Spectrometry data via Intelligent Clustering of Modulation Peaks.

Abstract In order to facilitate correlation calculation and matrix-based resolution in comprehensive two-dimensional gas chromatography - mass spectrometry (GC × GC-MS) data-set, an intelligent clustering of modulation peaks (ICMP) algorithm was developed in this paper. ICMP is start with the second -dimension (2D) peak restriction, then conducting the peak shape restriction in the first dimension (1D), finally end with the eigenvalues calculation against mass spectra in moving sub-windows. After this three-tier restriction, multi-component spectral correlative chromatography (MSCC) was applied in peak clustering result from a row-wise augmented "two-dimension (2D) slice" set. Then the component similarities and differences were distinguished rapidly/ accurately in chemical fingerprints from ChaiHu Shugan San and Cyperus rotundus. Faced with co-eluted phenomenon, matrix-based resolution was made in the representative sub-matrices that have been locked in ICMP procedure. From the example data shows that ICMP- multivariate curve resolution (MCR) can served as a good complement to (non) trilinear decomposition. To summarize, the GC × GC data-structure can be simplified to facilitate MSCC or MCR operation in fingerprints from herbal or biological samples.