A comparison of models used for calculation of RFLP pattern frequencies

In recent years the application of DNA typing information to criminal investigations has gained widespread acceptance. The primary method currently in use relies on length variation of DNA restriction fragments between individuals. These variations are identified using variable number tandem repeat (VNTR) DNA probes. As this technology becomes more widely used, it is crucial that scientifically valid methods of interpreting the significance of a DNA typing result be adopted. The method chosen should not only give a reliable approximation of the statistical likelihood of a particular RFLP pattern occurring, but should also be easy to present and for the court to understand. In this manuscript five methods of calculating a frequency of occurrence of a RFLP pattern will be presented (fixed bin genotype, floating bin phenotype, floating bin genotype, National Research Council (NRC) method using fixed bins and the NRC method using floating bins). The calculations discussed here demonstrated that the fixed bin genotype method produces a frequency very similar to that obtained from floating bin phenotypes. In addition, regardless of the method chosen or the database size, the frequency of any particular banding pattern in the population over several loci was found to be very rare.