EasyLCMS: an asynchronous web application for the automated quantification of LC-MS data

Abstract

Background: Downstream applications in metabolomics, as well as mathematical modelling, require data in a quantitative format, which may also necessitate the automated and simultaneous quantification of numerous metabolites. Although numerous applications have been previously developed for metabolomics data handling, automated calibration and calculation of the concentrations in terms of μmol have not been carried out. Moreover, most of the metabolomics applications are designed for GC-MS, and would not be suitable for LC-MS, since in LC, the deviation in the retention time is not linear, which is not taken into account in these applications. Moreover, only a few are web-based applications, which could improve stand-alone software in terms of compatibility, sharing capabilities and hardware requirements, even though a strong bandwidth is required. Furthermore, none of these incorporate asynchronous communication to allow real-time interaction with pre-processed results.

Findings: Here, we present EasyLCMS (http://www.easylcms.es/), a new application for automated quantification which was validated using more than 1000 concentration comparisons in real samples with manual operation. The results showed that only 1% of the quantifications presented a relative error higher than 15%. Using clustering analysis, the metabolites with the highest relative error distributions were identified and studied to solve recurrent mistakes.

Conclusions: EasyLCMS is a new web application designed to quantify numerous metabolites, simultaneously integrating LC distortions and asynchronous web technology to present a visual interface with dynamic interaction which allows checking and correction of LC-MS raw data pre-processing results. Moreover, quantified data obtained with EasyLCMS are fully compatible with numerous downstream applications, as well as for mathematical modelling in the systems biology field.

Figure 1

Number of publications related to analytical platforms used in targeted metabolomics. Number of publications related with quantification in the metabolomics area. Search criteria in ISI were ‘mass spectrometry’ AND quant* AND metab* AND ‘liquid chromatography’ for LC-MS, ‘gas chromatography’ for GC-MS or ‘capillary electrophoresis’ for CE-MS. For nuclear magnetic resonance, the term ‘mass spectrometry’ was substituted for ‘NMR’.

Figure 2

Structural overview of the EasyLCMS web application. EasyLCMS consists of three interacting layers: (i) web-interface, (ii) web-server processing modules and (iii) SQL-database. Details can be found in the EasyLCMS application tutorial.

Figure 3

Screenshots of the EasyLCMS web application. The quantification process is completely integrated and automated in the EasyLCMS platform. However, manual intervention is allowed at all steps to check and correct automatic results when needed.

Figure 4

Overview of the EasyLCMS data workflow. The quantification workflow schema is represented. Detailed instructions are provided in the main text and EasyLCMS application tutorial.

Figure 5

EasyLCMS performance using calibration standards. Calibration standard areas obtained using Chemstation software (x-axis) compared with areas obtained using EasyLCMS (y-axis) in all the experiments performed to validate EasyLCMS (see Supplemental Material). Linear regressions (red lines) and regression coefficients (R²) are represented for the 29 metabolites analysed.

Figure 6

Two-dimensional clustering representation of relative errors comparing EasyLCMS and Chemstation software with an internal standard. Twenty-one intracellular amino acids were quantified in samples from six different batch reactors of the human CCL-159 cell line (numbers 1–6), which were obtained with three different extraction procedures (ACN, MeOH and Chloro). See Supplemental Material for details.

Figure 7

Two-dimensional clustering representation of relative errors comparing EasyLCMS and Chemstation software without an internal standard. Twenty-nine metabolites were quantified in samples from two different human cell lines, CCL-159 (S) and CCL-159R (R), which were cultured in batch reactors. Samples were harvested at six different times (0, 23, 46, 71, 95 and 116 h) for extracellular (metabolite names include ‘ex’) as well as intracellular content. See Supplemental Material for details