Data Preprocessing Method for Liquid Chromatography–Mass Spectrometry Based Metabolomics

Abstract

A set of data pre-processing algorithms for peak detection and peak list alignment are reported for analysis of LC-MS based metabolomics data. For spectrum deconvolution, peak picking is achieved at selected ion chromatogram (XIC) level. To estimate and remove the noise in XICs, each XIC is first segmented into several peak groups based on the continuity of scan number, and the noise level is estimated by all the XIC signals, except the regions potentially with presence of metabolite ion peaks. After removing noise, the peaks of molecular ions are detected using both the first and the second derivatives, followed by an efficient exponentially modified Gaussian-based peak deconvolution method for peak fitting. A two-stage alignment algorithm is also developed, where the retention times of all peaks are first transferred into z-score domain and the peaks are aligned based on the measure of their mixture scores after retention time correction using a partial linear regression. Analysis of a set of spike-in LC-MS data from three groups of samples containing 16 metabolite standards mixed with metabolite extract from mouse livers, demonstrates that the developed data pre-processing methods performs better than two of the existing popular data analysis packages, MZmine2.6 and XCMS2, for peak picking, peak list alignment and quantification.

Untargeted LC–MS Data Preprocessing in Metabolomics

“…Soon after, a number of methods and software have been developed to circumvent such problem, such as COW, and MarkerLynx (Waters). XCMS 2 can align all samples with or without retention time correction [17]. Another method has been published to be able to perform bulk non-linear retention time correction in population-scale untargeted metabolomics data based on landmark features such as internal standards and invariant ubiquitous peaks in every sample [15].…”

“…Above all, the performance of untargeted metabolomics data processing software is reflected by the number of detected peaks, peak location (retention time, m/z value), peak alignment, peak area, which are the consideration for any researcher or company in developing data preprocess software [17].…”

“…In addition, the presence of numerous peaks from isotopes, adducts, and in-source degradation products, unexpected metabolites, and contaminants further complicate processing of the raw data [16]. Data preprocessing, including accurate spectrum deconvolution and peak alignment algorithms can significantly affect the outcome of downstream data analysis [17], and inappropriate treatment can lead to false discoveries.…”

“…The MatLAB-based software developed for untargeted lipidomics study is probably the first untargeted MS-based metabolomics tool, which used [17]. These tools are powerful in peak picking, deconvolution, and alignment.…”

Chemoselective reactions and their applications in metabolite isolation and analysis.

“…Such indiscriminate analyses result in highly complex and large data sets that require specialized software for analysis [9,10]. Although the data collection and analysis can be cumbersome for untargeted metabolomics, this approach leads to discovery of uncharacterized molecules and discovery of new pathways.…”