Compare LCMS - Extracting Component Differences in Similar LC/MS Data Sets

Antony Williams¹, Vitaly Lashin¹, Ilya Troisky¹, Willem Windig², William Nichols², Mary Goodberlet², William Smith^2
1 ACD/Labs, Toronto, Canada
²Eastman Kodak Company, Rochester, NY, USA

Abstract

Introduction

Scientists at Eastman Kodak previously reported the development of a component detection algorithm, CODA, for noise and background reduction in LC/MS chromatograms [1]. This was achieved using a variable selection procedure that selects mass chromatograms with low noise and low background contribution, then recombines the mass chromatograms to provide a reduced total ion chromatogram or optionally, overlay plots. Recently a second algorithm, CompareLCMS [2], was developed to distinguish dissimilarities in CODA reduced chromatograms permitting the extraction of component differences in similar LC/MS data sets.

While Windig's programs have proven effective for automating the processing and visualization of LC/MS data files, the targeted applications required program distribution to site wide desktops, a user-friendly interface, and file format support for a variety of MS vendors. These needs were achieved through collaboration with Advanced Chemistry Development, ACD/Labs. Implementation of the integrated Kodak and ACD/Labs programs consisted of three components:

• automated data processing for walk-up LC/MS users,
• visualization for comparative LC/MS studies, MS expert users, and
• the simultaneous reporting and capture of structure, spectra and sample information using Kodak's web based LIMS system, WIMS-DB [3], linked with ACD/Labs files.

These integrated tools are intended to reduce the data-handling bottleneck in LC/MS operations.

Results and Discussion

Synthetic chemists use LC/MS to follow synthetic reactions, both single and parallel, for structure and synthetic process discovery and optimization. Samples are analyzed on a PE-Sciex API-150 EX single quadrupole mass spectrometer with user selected HPLC methods, defined by chemistry, and automated column selection. The resulting data files are processed using CODA and printed in the walk-up laboratory to provide mass chromatograms and mass versus HPLC retention-time tables. The file is also available at the synthetic chemists desk top for processing, CODA, and viewing in ACD/MS manager (Figure 1). The information provides MW confirmation of reaction components and peak purity information for subsequent walk-up HPLC assays.

Figure 1. The synthetic chemist accesses the walk-up data at their desk top using ACD/MS Processor to view the resulting CODA plots, (peak purity), and spectral data, (MW information).

In contrast, the MS expert users frequently use LC/MS for comparative studies of two or more related samples and subsequent characterization of component differences. Data files from these studies are processed at the desktop using ACD/Labs with CompareLCMS, (Figure 2), to distinguish component differences and display the associated spectral data. This function eliminates the need for MS analysts to engage in the reiterative process of generating ion extracted plots, comparisons, and subsequent background subtraction, and spectral averaging.

Figure 2. Illustrates a comparative study using CompareLCMS in synthetic route discovery for commercialization. A synthetic sequence was run twice using methanol and acetone respectively to determine yield and impurity profile information.

Once the expert user has selected a spectrum, via CODA or CompareLCMS, for interpretation the proposed structure is attached and structure-spectral assignments are determined. The spectral files, with structure, assignments, annotation, and associated tables are linked to WIMS-DB. The ACD/Labs report editor captures the information as a report, linked to WIMS-DB, and a WIMS-DB report link is sent by email to the client for viewing at the desktop.

The viewed or retrieved WIMS-DB reports contain links to spectral files which, when activated, launch the ACD/viewer and display the "live" spectra-structure information. Activation of the report link launches ACD/ChemSketch displaying a printable report containing structure, spectra, and sample information. Data retrieval is accomplished through WIMS-DB using text queries such as sample information, and through ACD/Labs databases using structure, substructure, or spectral searching.

Conclusions

CODA provides automated data reduction for R&D synthetic chemist using walk-up LC/MS for routine peak purity and MW information needs. The walk-up automation has proven robust for batch samples from parallel synthesis as well as single-reaction experiments.

CompareLCMS is robust and accurate in comparative LC/MS studies including: formulation stability, synthetic discovery & optimization, reverse engineering, metabolite identification, and fundamental studies involving chemical interactions.

The collaboration with ACD/Labs enabled site-wide distribution of CODA for walk-up LC/MS users and for client report viewing through WIMS-DB. The combination of ACD/Labs and WIMS-DB captures our structure, spectra, and sample related information in concert with reporting.

The tools and processes described in this poster address the data-handling bottleneck in our LC/MS operations. Based on demonstrated efficiency gains, full implementation of this strategy will permit continued growth in our LC/MS applications using existing resources.

References

1. W. Windig, J. M. Phalp, A. W. Payne, A Noise and Background Reduction Method for Component Detection in Liquid Chromatography/Mass Spectrometry, Anal. Chem., 68, 1996, 3602-3606
2. W. Windig, W.F. Smith, W.F. Nichols, Fast Interpretation of Complex LC/MS Data Using Chemometrics, Anal. Chim. Act. 2001; 446: 465-474.
3. D. E. Brown, A. J. Williams, D. McLaughlin, Web-based Information Management System, Trends Anal. Chem., 16(7), 1997, 370-380

Download the poster (PDF file, 1.38 Mb).