Feature detection and alignment of hyphenated chromatographic–mass spectrometric data

Åberg, K. Magnus; Torgrip, Ralf J. O.; Kolmert, Johan; Schuppe-Koistinen, Ina; Lindberg, Johan

doi:10.1016/j.chroma.2008.03.033

Cited by 70 publications

(59 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In preparation for the comprehensive statistical analysis, mutant and wildtype strains were grown in small-scale fermentation in quadruplicate, replicate extracts were analyzed by LC-HRMS, and data were pretreated by using a compoundfinding algorithm, resulting in the definition of > 1000 molecular features per sample. [17][18][19] In order to identify molecular features specifically missing in culture extracts from DK1622 mutant strains, we applied principal-component analysis (PCA) to the preprocessed LC-MS datasets (Figure 1). PCA reduces the dimensionality of a multivariate dataset, while retaining the relevant information in terms of variance.…”

mentioning

confidence: 99%

Myxoprincomide: A Natural Product from Myxococcus xanthus Discovered by Comprehensive Analysis of the Secondary Metabolome

Cortina¹,

Krug²,

Plaza³

et al. 2011

Angew Chem Int Ed

110

View full text Add to dashboard Cite

mentioning

confidence: 99%

Myxoprincomide: A Natural Product from Myxococcus xanthus Discovered by Comprehensive Analysis of the Secondary Metabolome

Cortina¹,

Krug²,

Plaza³

et al. 2011

Angew Chem Int Ed

110

View full text Add to dashboard Cite

“…Here, the TracMass algorithm [21] feature detects by the means of individually scan-by-scan tracking Kalman filters for each occurring peak, analogously to how airplanes are tracked in radar data. This method operates on centroided MS data and effectively detects consistent features by support of both the mass-and chromatography dimension.…”

Section: Peak Detection In 2d Datamentioning

confidence: 99%

“…Panes A-C depicts portions of NMR, LC-MS and CE-MS data respectively. Sub-panes (a) shows top and bottom feature vectors from sub-panes (c), feature vectors (spectra, Pure Ion Chromatograms [21] (PIC) and Total Ion Chromatograms (TIC)). Sub-panes (b) shows unsorted heat maps of feature vectors from many samples, colour indicates intensity.…”

Section: Peak or Feature Detectionmentioning

confidence: 99%

Warping and alignment technologies for inter-sample feature correspondence in 1D H-NMR, chromatography-, and capillary electrophoresis-mass spectrometry data

2010

Self Cite

View full text Add to dashboard Cite

Comprehensive analysis is an emerging and interesting mode of analyzing biological samples. The hypothesis is that by measuring the abundance of as many analytes as possible in samples of biological origin, new knowledge about biological regulation and pathways and can be gained. This hypothesis is one of the cornerstones of the -omics sciences. The comprehensive measurements are usually performed by H-NMR or hyphenated mass spectrometry of almost intact biofluids and tissues. The data generated from these measurements are very complex, containing thousands of analyte signals, necessitating automatic information extraction. The success of this information extraction is dependent on many steps and a most crucial step is solving the inter-sample correspondence problem, i.e. that analytes from different samples are unsynchronized in the measured data. This problem hampers the data-analysis pipeline by introducing errors in the resulting extracted abundance data. The remedy is usually tedious manual inspection for analyte synchronization. In this paper, we address automated methods for addressing the inter-sample correspondence problem when analyzing data from comprehensive measurements, i.e. synchronization methods. We here present a review based on methods taxonomy, complexity, and an overview of current solution technologies available.

show abstract

“…Peaks from the same metabolite may shift in the RT dimension over the course of, and between, analytical runs, and these must be matched together for the final data table. Many algorithms have been proposed [8][9][10] and, although there is much room for improvement, the results are usually good enough to be useful for further analysis. Finally, the intensity of each peak must be estimated, usually by a simple integration of the ion count across the RT dimension, possibly with additional noise reduction, smoothing and/or baseline correction steps.…”

Section: Lc-msmentioning

confidence: 99%

Statistical Data Analysis in Metabolomics

Ebbels

2011

Handbook of Statistical Systems Biology

View full text Add to dashboard Cite

Metabolomics, also known as metabonomics or metabolic profiling [1,2], is the study of the complement of small molecule metabolites present in cells, biofluids and tissues. Metabolites (biomolecules with masses below about 1500 Da) are the key molecules that sustain life, providing the energy and raw materials of the cell. As such, any disregulation or malfunction of the system, due for example to disease or stress, will usually cause changes in metabolite levels, which can be monitored by spectroscopic techniques. Metabolites form one component of the multilevel-systems approach to biology: they interact dynamically with macromolecules in a host of other levels and multiple spatial scales. For example, metabolite levels both regulate and are regulated by genes through the transcriptional and protein levels of biomolecular organisation, and in some cases may be instrumental in cross-species interactions, for example in symbiosis or parasitism. In contrast to proteins and nucleic acids, the metabolic profile is directly sensitive to the environment of the system, for example being strongly influenced by drugs and diet.In metabolomics, the aim is to obtain a global profile of metabolite levels with as little bias as possible, which is the so-called untargeted approach. This poses a severe challenge to analytical chemical procedures that are traditionally optimised for specific target molecules or chemical classes. In order to capture the widest set of metabolites, multiple chemical analytical technologies must be employed. Each technology results in large and complex datasets requiring extensive processing and statistical modelling to draw biological inferences. In this chapter we present some of the most widely used statistical techniques for analysing metabolomic data. We highlight the challenges, including those still to be addressed, and begin by describing the technologies used and how they influence the statistical characteristics of the data. In the first part of the chapter, we discuss the range of analytical tools used to acquire metabolic profiles and the associated characteristics of the data. In the second part, we focus on statistical approaches that can be applied independently of the analytical Handbook of Statistical Systems Biology, First Edition. Edited

show abstract

Feature detection and alignment of hyphenated chromatographic–mass spectrometric data

Cited by 70 publications

References 14 publications

Myxoprincomide: A Natural Product from Myxococcus xanthus Discovered by Comprehensive Analysis of the Secondary Metabolome

Myxoprincomide: A Natural Product from Myxococcus xanthus Discovered by Comprehensive Analysis of the Secondary Metabolome

Warping and alignment technologies for inter-sample feature correspondence in 1D H-NMR, chromatography-, and capillary electrophoresis-mass spectrometry data

Statistical Data Analysis in Metabolomics

Contact Info

Product

Resources

About