Lauren McIntyre and B.S. Weir
North Carolina State University
The issue of independence of alleles within VNTR loci has been addressed with intraclass correlations on fragment lengths, and by various tests on discretized or binned lengths. As the data are essentially continuous, the former approach has some appeal although the latter is more relevant to the way in which the data are used in identity calculations. As part of an investigation of alternatives to match/binning strategies, we have used kernel density estimation to estimate continuous genotypic distributions. We have compared several testing strategies to compare these two-dimensional genotypic densities to the corresponding products of one-dimensional allelic densities. We will show these results and compare them to the findings from exact tests on discretized data.
Go to proceedings home page