An iterative block-shifting approach to retention time alignment that preserves the shape and area of gas chromatography-mass spectrometry peaks

Abstract

Background: Metabolomics, petroleum and biodiesel chemistry, biomarker discovery, and other fields which rely on high-resolution profiling of complex chemical mixtures generate datasets which contain millions of detector intensity readings, each uniquely addressed along dimensions of time (e.g., retention time of chemicals on a chromatographic column), a spectral value (e.g., mass-to-charge ratio of ions derived from chemicals), and the analytical run number. They also must rely on data preprocessing techniques. In particular, inter-run variance in the retention time of chemical species poses a significant hurdle that must be cleared before feature extraction, data reduction, and knowledge discovery can ensue. Alignment methods, for calibrating retention reportedly (and in our experience) can misalign matching chemicals, falsely align distinct ones, be unduly sensitive to chosen values of input parameters, and result in distortions of peak shape and area.

A chemometric study of chromatograms of tea extracts by correlation optimization warping in conjunction with PCA, support vector machines and random forest data modeling

“…Robust alignment is required to improve the quality of chromatography data but also in order to determine which type of data analysis should be considered [12]. The study of alignment algorithms has increasingly become a separate subject in this field [13][14][15].…”

A novel strategy for quantitative analysis of the formulated complex system using chromatographic fingerprints combined with some chemometric techniques

“…Whereas, those frequently used one-dimensional wrapping algorithms [32][33][34][35][36][37][38][39][40][41] cannot make full use of the second dimensional information provided by multi-channel detector. Accordingly, several alignment methods for hyphenated chromatographic data [42][43][44][45][46] have emerged in recent years.…”

Unsupervised adaptive filter for baseline thresholding and elimination in liquid chromatography-mass spectrometry via approximation of the standard deviation of baseline distribution in retention time domain

“…Strong peak candidates allow the alignment to have additional flexibility [8,20]. Robust peak detectors require advanced analysis, e.g., noise filtration, baseline subtraction, pattern recognition, or curve fitting [20][21][22]. Noise additions are caused not only by random errors (random noise) but also by an influence of baseline from the liquid chromatography.…”