A simple method for establishing concordance between short-tandem-repeat allele frequency databases

Abstract

BACKGROUND: Current paternity and forensic accreditation standards do not require concordance to be established between short-tandem-repeat allele frequency databases representing the same population. The current statistical methods for evaluating databases do not establish concordance. Although acceptable under current forensic statistical methods, databases representing the same population may have sufficient variation to influence the outcome of a nondirect relationship testing result (i.e., siblingship). Hence there is a need for a quantitative method to determine concordance between databases. STUDY DESIGN AND METHODS: Local allele frequency databases were generated for major US ethnic groups. Statistical analysis was performed as recommended by international forensic standards. A new method was developed and used for evaluating concordance between the locally developed and published databases. Smaller deviation values signify greater concordance between compared databases. The results were quantitatively confirmed against data obtained from a multidimensional scaling analysis system (SPSS, SPSS, Inc.). RESULTS: The locally developed database deviated from the mean of the published data by approximately 0.1073 for Caucasians, 0.1341 for US Hispanics, and 0.1287 for African Americans. Upon pairwise comparison, the published databases deviated from one another by 0.1210 in US Caucasians, 0.1457 in US Hispanics, and 0.1228 in African Americans. The local database was observed to be more concordant to published databases than some of the published data to themselves. CONCLUSION: The developed database was shown to be concordant with four previously published databases using the new method. The results were confirmed by comparison to qualitative data obtained from the multidimensional scaling analysis system. © 2010 American Association of Blood Banks.