Skip to main content
ARS Home » Northeast Area » Beltsville, Maryland (BHNRC) » Beltsville Human Nutrition Research Center » Food Composition and Methods Development Laboratory » Research » Publications at this Location » Publication #347611
Title: A Note on the Use of Workstation Software Programs for Quantification

Author
item Byrdwell, W Craig

Submitted to: Journal of Liquid Chromatography and Related Technologies
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 6/26/2019
Publication Date: 7/15/2019
Citation: Byrdwell, W.C. 2019. A Note on the Use of Workstation Software Programs for Quantification. Journal of Liquid Chromatography and Related Technologies. 4217-18:570-574. https://doi.org/10.1080/10826076.2019.1637350.
DOI: https://doi.org/10.1080/10826076.2019.1637350

Interpretive Summary: These days, instrument manufacturers are working to make their instruments as user-friendly as possible. Users often want to use a "black box" approach that provides desired values for quantification with the least user intervention possible. To accommodate users, makers of instrument workstation software programs sometimes use approaches to quantification that are not statistically valid and are not suitable for publication. However, the users may be unaware of the approaches to quantification that the instruments employ. This report demonstrates the shortcuts that some instruments incorporate into their software to give users better calibration parameters to report, even though these are not statistically rigorous. Results are provided to prove that users need to verify the calculations used by instrument workstations to make sure their results are publishable.

Technical Abstract: Most chromatographic and mass spectrometric instruments include workstations that allow data processing and target compound quantification. However, the software used for quantification does not all use common approaches for statistical treatment of data. Presented here is a brief description of three commonly used workstation software packages (WSPs) and the degree to which they obey widely accepted approaches to statistical treatment of data. It was found that the Thermo Fisher Scientific software packages, Xcalibur and TraceFinder, provided calibration lines and accompanying parameters that were similar or identical to those obtained by generic treatment using the Microsoft Excel ‘linest()’ function, whereas the Agilent OpenLab ChemStation software did not. It is recommended to always perform calculations manually via spreadsheets to allow better statistical treatment of data and to confirm whether the WSP used employs the commonly accepted approach to linear calibration.