Nonparametric measures of correlations of DNA fragment lengths within and between variable number of tandem repeat (VNTR) loci are proposed to test the hypothesis of random association of allele sizes at VNTR loci. Transformations of these nonparametric correlation measures are suggested to detect deviations of their null expectations caused by population subdivision and errors of measurement of VNTR fragment lengths. Analytic and permutation- based computer simulation studies are performed to show that under the hypothesis of independence of allele sizes the transformed correlation measures are normally distributed, irrespective of the VNTR fragment size distribution in the population even when the number of individuals samples is as low as 100. Power calculations are performed to establish that the current population data on six VNTR loci in the US Hispanic sample are in accordance with the hypothesis of random association of allele sizes within and between loci. Implications of these results in the context of forensic use of DNA typing are also discussed.
|Number of pages||9|
|State||Published - 1993|