Abstract
Multiphase designs are an effective technique to assess all sources of variation in an experiment; however they have not yet been widely applied to analytical chemistry. In this study a multiphase design was used to evaluate sources of error from the field and laboratory phases of an experiment involving canola growing, oil extraction, sample preparation, and analysis using high performance liquid chromatography coupled to diode array detection and tandem mass spectrometry (HPLC-DAD-MS/MS). Several classes of bioactive compounds ‒ tocopherols, carotenoids and sterols ‒ were measured in the canola (Brassica napus) oil from 64 different genotypes to be assessed for varietal and environmental influence. Other factors which might contribute to error were identified, incorporated into the design, and their associated error calculated and accounted for. The field and HPLC phases were the largest contributors to total error, with only small influences from the extraction and preparation phases. Likelihood ratio testing of the nested multiphase models proved that high precision was achieved by use of the multiphase design, and identified possible improvements for future laboratory work. In the first reported application of a multiphase design to multi-stage laboratory analyses, the design was shown to offer considerable advantages over traditional approaches particularly in reducing total sample number, time and cost of analysis, as well as more comprehensive monitoring of experimental error.
Original language | English |
---|---|
Pages (from-to) | 55-64 |
Number of pages | 10 |
Journal | Chemometrics and Intelligent Laboratory Systems |
Volume | 162 |
Early online date | Jan 2017 |
DOIs | |
Publication status | Published - Mar 2017 |