Measurement issues associated with quantitative molecular biology analysis of complex food matrices for the detection of food fraud.

Following a report on a significant amount of horse DNA being detected in a beef burger product on sale to the public at a UK supermarket in early 2013, the Elliott report was published in 2014 and contained a list of recommendations for helping ensure food integrity. One of the recommendations included improving laboratory testing capacity and capability to ensure a harmonised approach for testing for food authenticity. Molecular biologists have developed exquisitely sensitive methods based on the polymerase chain reaction (PCR) or mass spectrometry for detecting the presence of particular nucleic acid or peptide/protein sequences. These methods have been shown to be specific and sensitive in terms of lower limits of applicability, but they are largely qualitative in nature. Historically, the conversion of these qualitative techniques into reliable quantitative methods has been beset with problems even when used on relatively simple sample matrices. When the methods are applied to complex sample matrices, as found in many foods, the problems are magnified resulting in a high measurement uncertainty associated with the result which may mean that the assay is not fit for purpose. However, recent advances in the technology and the understanding of molecular biology approaches have further given rise to the re-assessment of these methods for their quantitative potential. This review focuses on important issues for consideration when validating a molecular biology assay and the various factors that can impact on the measurement uncertainty of a result associated with molecular biology approaches used in detection of food fraud, with a particular focus on quantitative PCR-based and proteomics assays.

Improved evaluation of measurement uncertainty from sampling by inclusion of between-sampler bias using sampling proficiency testing.

A realistic estimate of the uncertainty of a measurement result is essential for its reliable interpretation. Recent methods for such estimation include the contribution to uncertainty from the sampling process, but they only include the random and not the systematic effects. Sampling Proficiency Tests (SPTs) have been used previously to assess the performance of samplers, but the results can also be used to evaluate measurement uncertainty, including the systematic effects. A new SPT conducted on the determination of moisture in fresh butter is used to exemplify how SPT results can be used not only to score samplers but also to estimate uncertainty. The comparison between uncertainty evaluated within- and between-samplers is used to demonstrate that sampling bias is causing the estimates of expanded relative uncertainty to rise by over a factor of two (from 0.39% to 0.87%) in this case. General criteria are given for the experimental design and the sampling target that are required to apply this approach to measurements on any material.

Scoring in genetically modified organism proficiency tests based on log-transformed results.

The study considers data from 2 UK-based proficiency schemes and includes data from a total of 29 rounds and 43 test materials over a period of 3 years. The results from the 2 schemes are similar and reinforce each other. The amplification process used in quantitative polymerase chain reaction determinations predicts a mixture of normal, binomial, and lognormal distributions dominated by the latter 2. As predicted, the study results consistently follow a positively skewed distribution. Log-transformation prior to calculating z-scores is effective in establishing near-symmetric distributions that are sufficiently close to normal to justify interpretation on the basis of the normal distribution.

Determination of deoxynivalenol in cereals and cereal products by immunoaffinity column cleanup with liquid chromatography: interlaboratory study.

An interlaboratory study was performed on behalf of the UK Food Standards Agency to evaluate the effectiveness of an immunoaffinity column cleanup liquid chromatographic (LC) method for the determination of deoxynivalenol in a variety of cereals and cereal products at proposed European regulatory limits. The test portion was extracted with water. The sample extract was filtered a applied to an immunoaffinity column. After being washed with water, the deoxynivalenol was eluted with acetonitrile or methanol. Deoxynivalenol was quantitated by reversed-phase LC with UV determination. Samples of artificially contaminated wheat-flour, rice flour, oat flour, polenta, and wheat based breakfast cereal, naturally contaminated wheat flour, and blank (very low level) samples of each matrix were sent to 13 collaborators in 7 European countries. Participants were asked to spike test portions of all samples at a range of deoxynivalenol concentrations equivalent to 200-2000 ng/g deoxynivalenol. Average recoveries ranged from 78 to 87%. Based on results for 6 artificially contaminated samples (blind duplicates), the relative standard deviation for repeatability (RSDr) ranged from 3.1 to 14.1%, and the relative standard deviation for reproducibility (RSDR) ranged from 11.5 to 26.3%. The method showed acceptable within-laboratory and between-laboratory precision for all 5 matrixes, as evidenced by HorRat values < 1.3.

Determination of zearalenone in barley, maize and wheat flour, polenta, and maize-based baby food by immunoaffinity column cleanup with liquid chromatography: interlaboratory study.

An interlaboratory study was performed on behalf of the UK Food Standards Agency to evaluate the effectiveness of an affinity column cleanup liquid chromatography (LC) method for the determination of zearalenone (ZON) in a variety of cereals and cereal products at proposed European regulatory limits. The test portion is extracted with acetonitrile:water. The sample extract is filtered, diluted, and applied to an affinity column. The column is washed, and ZON is eluted with acetonitrile. ZON is quantified by reversed-phase LC with fluorescence detection. Barley, wheat and maize flours, polenta, and a maize-based baby food naturally contaminated, spiked, and blank (very low level) were sent to 28 collaborators in 9 European countries and 1 collaborator in New Zealand. Participants were asked to spike test portions of all samples at a ZON concentration equivalent to 100 microg/kg. Average recoveries ranged from 91-111%. Based on results for 4 artificially contaminated samples (blind duplicates) and 1 naturally contaminated sample (blind duplicate), the relative standard deviation for repeatability (RSDr) ranged from 6.9-35.8%, and the relative standard deviation for reproducibility (RSDR) ranged from 16.4-38.2%. The method showed acceptable within- and between-laboratory precision for all 5 matrixes, as evidenced by HorRat values <1.7.

Testing for bias between the Kjeldahl and Dumas methods for the determination of nitrogen in meat mixtures, by using data from a designed interlaboratory experiment.

Bias between the Dumas and the Kjeldahl methods for the determination of protein nitrogen in food was studied by conducting an interlaboratory study involving 40 laboratories and 20 different test materials. Biases were found to be small and statistically significant only for the chicken test materials, where a bias of 0.020±0.004% m/m was detected.

Evaluation of methods of analysis for dietary fibre using real foods and model foods.

RATIONALE AND OBJECTIVES: There are several rational and empirical methods for the measurement of dietary fibre and its components. A selection of these methods were evaluated by investigation of a range of real foods and model foods with added resistant starch (RS), non-starch polysaccharides (NSP) and resistant oligosaccharide (RO) ingredients. METHODS: A range of rational methods were applied in determining specific carbohydrate constituents: RS, NSP and RO, including fructans. For comparison, empirical methods AOAC 991.43 (2001.03) and AOAC 2009.01 were applied, based on determination of gravimetric residues for high molecular weight and size-exclusion HPLC analysis of the ethanol filtrate for low molecular weight components. RESULTS: In general there was agreement between different rational methods for the analysis of RS and fructans, though there were notable exceptions for some product types. Comparison of methods for total RS and those that only measure the RS3 fraction, from retrograded starch, indicated that RS3 was the only type present for most processed products. This also explains the similar results obtained by AOAC 991.43 (2001.03) and AOAC 2009.01, though the latter is intended to recover other RS types as well. For many products there was agreement between results obtained by rational and empirical methods, though there were exceptions and the reasons for these are discussed. CONCLUSION: Rational and empirical methods can both be used to determine dietary fibre in most situations. The information provided by rational methods is useful in identifying the specific carbohydrate constituents present in foods and can assist in determining whether added extracted and synthesised ingredients are ones that conform to the Codex and EU dietary fibre definition.

The duplicate method of uncertainty estimation: are eight targets enough?

This paper presents methods for calculating confidence intervals for estimates of sampling uncertainty (s(samp)) and analytical uncertainty (s(anal)) using the chi-squared distribution. These uncertainty estimates are derived from application of the duplicate method, which recommends a minimum of eight duplicate samples. The methods are applied to two case studies--moisture in butter and nitrate in lettuce. Use of the recommended minimum of eight duplicate samples is justified for both case studies as the confidence intervals calculated using greater than eight duplicates did not show any appreciable reduction in width. It is considered that eight duplicates provide estimates of uncertainty that are both acceptably accurate and cost effective.

Empirical versus modelling approaches to the estimation of measurement uncertainty caused by primary sampling.

Measurement uncertainty is a vital issue within analytical science. There are strong arguments that primary sampling should be considered the first and perhaps the most influential step in the measurement process. Increasingly, analytical laboratories are required to report measurement results to clients together with estimates of the uncertainty. Furthermore, these estimates can be used when pursuing regulation enforcement to decide whether a measured analyte concentration is above a threshold value. With its recognised importance in analytical measurement, the question arises of 'what is the most appropriate method to estimate the measurement uncertainty?'. Two broad methods for uncertainty estimation are identified, the modelling method and the empirical method. In modelling, the estimation of uncertainty involves the identification, quantification and summation (as variances) of each potential source of uncertainty. This approach has been applied to purely analytical systems, but becomes increasingly problematic in identifying all of such sources when it is applied to primary sampling. Applications of this methodology to sampling often utilise long-established theoretical models of sampling and adopt the assumption that a 'correct' sampling protocol will ensure a representative sample. The empirical approach to uncertainty estimation involves replicated measurements from either inter-organisational trials and/or internal method validation and quality control. A more simple method involves duplicating sampling and analysis, by one organisation, for a small proportion of the total number of samples. This has proven to be a suitable alternative to these often expensive and time-consuming trials, in routine surveillance and one-off surveys, especially where heterogeneity is the main source of uncertainty. A case study of aflatoxins in pistachio nuts is used to broadly demonstrate the strengths and weakness of the two methods of uncertainty estimation. The estimate of sampling uncertainty made using the modelling approach (136%, at 68% confidence) is six times larger than that found using the empirical approach (22.5%). The difficulty in establishing reliable estimates for the input variable for the modelling approach is thought to be the main cause of the discrepancy. The empirical approach to uncertainty estimation, with the automatic inclusion of sampling within the uncertainty statement, is recognised as generally the most practical procedure, providing the more reliable estimates. The modelling approach is also shown to have a useful role, especially in choosing strategies to change the sampling uncertainty, when required.

Two-stage application of the optimised uncertainty method: a practical assessment.

Uncertainty estimates from routine sampling and analytical procedures can be assessed as being fit for purpose using the optimised uncertainty (OU) method. The OU method recommends an optimal level of uncertainty that should be reached in order to minimise the expected financial loss, given a misclassification of a batch as a result of the uncertainty. Sampling theory can used as a predictive tool when a change in sampling uncertainty is recommended by the OU method. The OU methodology has been applied iteratively for the first time using a case study of wholesale butter and the determination of five quality indicators (moisture, fat, solids-not-fat (SNF), peroxide value (PV) and free fatty acid (FFA)). The sampling uncertainty (s(samp)) was found to be sub-optimal for moisture and PV determination, for 3-fold composite samples. A revised sampling protocol was devised using Gy's sampling theory. It was predicted that an increase in sample mass would reduce the sampling uncertainty to the optimal level, resulting in a saving in expectation of loss of over pounds 2000 per 20 tonne batch, when compared to current methods. Application of the optimal protocol did not however, achieve the desired reduction in s(samp) due to limitations in sampling theory. The OU methodology proved to be a useful tool in identifying broad weaknesses within a routine protocol and assessing fitness for purpose. However, the successful routine application of sampling theory, as part of the optimisation process, requires substantial prior knowledge of the sampling target.

Optimising uncertainty in physical sample preparation.

Uncertainty associated with the result of a measurement can be dominated by the physical sample preparation stage of the measurement process. In view of this, the Optimised Uncertainty (OU) methodology has been further developed to allow the optimisation of the uncertainty from this source, in addition to that from the primary sampling and the subsequent chemical analysis. This new methodology for the optimisation of physical sample preparation uncertainty (u(prep), estimated as s(prep)) is applied for the first time, to a case study of myclobutanil in retail strawberries. An increase in expenditure (+7865%) on the preparatory process was advised in order to reduce the s(prep) by the 69% recommended. This reduction is desirable given the predicted overall saving, under optimised conditions, of 33,000 pounds Sterling per batch. This new methodology has been shown to provide guidance on the appropriate distribution of resources between the three principle stages of a measurement process, including physical sample preparation.

A comparison of the Kjeldahl and Dumas methods for the determination of protein in foods, using data from a proficiency testing scheme.

Both the Kjeldahl and the Dumas methods for the determination of protein in foodstuffs are currently in use, but the empirical nitrogen factors used to convert the determined nitrogen content to protein content are based on the Kjeldahl method alone. Non-equivalence between the two methods could therefore result in some laboratories reporting an incorrect protein content. We report here a study using data accumulated over several years in the results of a proficiency testing scheme. On average the Dumas method provided results that were relatively higher by about 1.4% than the Kjeldahl method, but the difference between the methods depended on the type of foodstuff. The methodology of looking for bias between analytical methods is critically discussed.

A pilot study of routine quality control of sampling by the SAD method, applied to packaged and bulk foods.

A recently proposed method of looking at sampling uncertainty has been tested by its application to the sampling and analysis of several types of food and an animal feedstuff. In this 'SAD' method, increments comprising the conventional sample (that is, collected in the fashion prescribed by the standard sampling protocol) are allocated to either of two equal sized 'splits', which are prepared and analysed separately. The absolute difference between the analytical results for the two splits (the split absolute difference, or SAD) is plotted on a one-sided control chart. A non-compliance indicates that the combined uncertainty of sampling or analysis is larger than expected and the result of the measurement (the mean of the two split results) is possibly not fit for purpose. In addition, the SAD results give rise to a rugged estimate the uncertainty associated with the sampling protocol, often a major part of the total measurement uncertainty.

Determination of isotopically labelled monoesterphthalates in urine by high performance liquid chromatography-mass spectrometry.

A method of analysis for monoesters of phthalic acid ('monoesterphthalates') in human urine has been developed. The method was needed to determine the hydrolysis and excretion efficiency of isotopically-labelled phthalate diesters ('phthalates') when they were fed to volunteers as part of a biomarker study to estimate total exposure to phthalates. The targeted substances were 13C-monobutylphthalate (MBP), 2H4-monobutylphthalate (MBP), 2H4-monobenzylphthalate (MBeP), 13C-monocyclohexylphthalate (MCHP), 13C-monoethylhexylphthalate (MEHP), and 13C-monoisodecylphthalate (MIDP). The monoesters in urine were deconjugated enzymatically, extracted into solvent, and then determined by high performance liquid chromatography-mass spectrometry (LC-MS) using atmospheric pressure chemical ionisation in the negative ion mode. The limits of determination were 10 ng ml(-1) for MBP, MCHP, MBeP and MEHP, and 40 ng ml(-1) for MIDP. The recovery from urine spiked at 100 ng ml(-1) was in the range from 70 to 85% except for MIDP which was lower at 55%. The between-batch reproducibility of the analysis was in the range 8 to 17% (n = 6 batches on separate days).