No abstract

Extraction of qualitative and quantitative information from large numbers of analytical signals is difficult with drifted baselines, particularly in multivariate analysis. Baseline drift obscures and "fuzzies" signals, and even deteriorates analytical results. In order to obtain accurate and clear results, some effective methods should be proposed and implemented to perform baseline correction before conducting further data analysis. However, most of the classic methods require user intervention or are prone to variability, especially with low signal-to-noise signals. In this study, a novel baseline correction algorithm based on quantile regression and iteratively reweighting strategy is proposed. This does not require user intervention and prior information, such as peak detection. The iteratively reweighting strategy iteratively changes weights of residuals between fitted baseline and original signals. After a series of tests and comparisons with several other popular methods, using various kinds of analytical signals, the proposed method is found to be fast, flexible, robust, and easy to use both in simulated and real datasets.

Metabolomics datasets generated by modern analytical instruments tend to be increasingly complex. In this study, a recent method named shrunken centroids regularized discriminant analysis (SCRDA) has been introduced and applied in the exploration of metabolomics dataset. It is a supervised method for variable selection, discriminant analysis and biomarker screening. By regularizing the estimate of the within‐class covariance matrix, SCRDA can deal with the singularity issue of linear discriminant analysis. Then a shrinkage estimator is applied to perform variable selection. The method presented is illustrated through the simulated datasets and three complex metabolomics datasets. Commonly used orthogonal partial least squares discriminant analysis and two other similar statistical methods, penalized linear discriminant analysis and nearest shrunken centroids, are used for comparisons. The results illustrate that SCRDA has some desirable abilities in variable selection, classification and prediction. Moreover, the biomarkers identified by SCRDA are further demonstrated to be in accordance with the biochemical research. It has been proved that SCRDA can be applied as a promising strategy in metabolomics. Copyright © 2014 John Wiley & Sons, Ltd.

The discovery of potential biomarkers that may be closely related to diseases is a major purpose of metabolomics data analysis. Hence, we expect to find some effective methods which can screen these potential biomarkers from large amounts of dataset. In this paper, we propose an effective strategy named sparse linear discriminant analysis (SLDA), which can perform classification and variable selection simultaneously to analyze complicated metabolomics datasets. Compared with two other approaches, i.e. partial least squares discriminant analysis (PLS-DA) and competitive adaptive reweighted sampling (CARS), SLDA obtains relatively better results and can identify some informative metabolites, which are proven to be consistent with those identified by biochemical studies. Furthermore, by building a model based on selected features, SLDA can be applied to high dimensional, small sample cases where linear discriminant analysis (LDA) fails to work. In summary, SLDA is a very useful method for exploring and processing metabolomics data.