WP4 (statistical data analysis)

Workpackage leader: Roy Macarthur, FERA


WP4 will provide statistical analysis and interpretation of the results produced during the validation of methods. The aim of the statistical analysis of the results produced during method validation is to provide an objective assessment of how methods perform. We will examine a number of different aspects of performance, but this assessment will boil down to two headline numbers per method:

  • an estimate of an upper limit for the false positive probability and
  • an estimate of an upper limit for the limit of detection of the method.

The false positive probability is the expected proportion of samples that don’t contain the target GMO but will still produce a positive result. The limit of detection is the smallest amount of GMO that will reliably be detected if it is tested. To put it another way, if we test a sample and get a negative result we can’t be sure that the sample contains 0% GMO but we can be pretty sure that it contains a less than the limit of detection. The most useful test methods have a low false positive probability and a low limit of detection, but sometimes a method which is only reliable in one direction can also be useful.

We will be basing our analysis on methods recently reported by Wehling [1] and Macarthur [2]. These are methods that are designed to deal with the validation of qualitative methods of analysis. They provide an estimate of the probability of detection and the range within which results produced by individual laboratories can be expected to lie when they use the method. This means that we are able to provide estimates of the performance of methods for detecting the presence of GMOs that are robust to the normal variation we can expect to see between different laboratories using the methods.

Analysis of a qualitative method of detection to estimate its limit of detection and false positive probability

1 Wehling P et al J AOAC Int. 2011 Jan-Feb;94(1):335-47.
2 Macarthur, R, von Holst, Anal. Methods, 2012,4, 2744-2754

Project GMOVal

The central aim of the project is to make available to the GMO laboratories community new validated real-time PCR screening methods for an improved GMO detection via the implementation of the matrix approach.