PubAg

Main content area

Automated Peak Extraction and Quantification in Chromatography with Multichannel Detectors

Author:
Furbo, Søren, Christensen, Jan H.
Source:
Analytical chemistry 2012 v.84 no.5 pp. 2211-2218
ISSN:
1520-6882
Subject:
detectors, factor analysis, gas chromatography-mass spectrometry, models, polycyclic aromatic hydrocarbons
Abstract:
Chromatography combined with spectral analyzers provides great separation powers. However, the amount of data provided makes analysis difficult and laborious. The automated peak extraction and quantification (APEQ) method is a novel method for targeted and untargeted analysis. The method extracts retention times (tR), spectra, and signal intensity from data from chromatography with spectral detectors. It is a tiered approach consisting of five tiers: (1) the chromatograms are split into sections. (2) Peaks are detected and their elution profiles, spectra, and signal intensities are approximated. (3) The number of important peaks in each section is estimated. (4) For each section, the peaks and background are modeled using parallel factor analysis (PARAFAC). Several PARAFAC models with a different number of factors are made. (5) For each section, the optimal model is selected. The optimal model will in most cases be the model with lowest dimensionality (i.e., fewest factors) that has a factor describing the baseline. The APEQ method has been used to analyze gas chromatography/mass spectrometry (GC/MS) full-scan chromatograms of 12 samples, each containing 30 polycyclic aromatic hydrocarbons (PAHs). It detected all 30 PAHs and 14 compounds not intentionally added. It extracted the tRs, mass spectra, and relative concentrations. The quantification was compared to integration of extracted ion chromatograms. It had comparable quality for peaks with little or no overlap and significantly better for peaks with overlap. The total analysis time with default parameters was less than 10 min on an ordinary laptop. By optimizing the parameters, all 30 PAHs could be quantified in 142 s.
Agid:
5291628