R.M. Fourney1, C.J. Fregeau1, J.H. Bowen1, K.L. Bowen1,
G.G. Shutlerr1, J.C. Elliott1, P.D. Roussy1 and G. R.
Carmody1,2
1Biology Services, Central Forensic Laboratory, Royal Canadian Mounted
Police, Ottawa, Ontario, Canada
2Dept. of Biology, Carleton University, Ottawa, Ontario, Canada
Fewer observed alleles and better match precision based on size measurement of STR loci, will potentially impact on the frequency estimation of the resulting combined multilocus genotype. Over the past three years, the RCMP has investigated the detection and classification of numerous STR loci using an automated fluorescent detection procedure (Applied Biosystems Division Perkin Elmer: 373 or 377 Sequencer). The fluorescent detection of these alleles using laser-excitation detection systems and molecular weight internal lane standards has increased our ability to quantitate and analyse amplified products. To date, this has proven to provide excellent precision and accuracy for reproducible allele designation based on size measurement calculated by the local Southern method using internal lane markers. The measurement of instrument precision and reproducibility of representative alleles from nine STR systems were determined using different internal molecular weight standards. The intragel and intergel precision was evaluated to be ± 0.5 bases and ± 0.75 bases respectively (ABD 373).
Designation of the STR alleles based on size determination has been shown to be valid and reliable but has uncovered interesting variation that potentially reflects the inherent sequence variation. Estimation of allele frequency using a floating bin approach and the measurement precision of nominal STR alleles to delimit the floating range or "bin" (± 3 standard deviations) provides a conservative estimation of genotype allele frequencies and compensates for potential biological sequence variants. In addition the frequencies of rare alleles are estimated using a minimum allele frequency ("floor") based on sample size according to Chakraborty (1992) and Weir (1992). This approach provides a statistically sound and conservative method of estimating the frequency of an STR genotype. Casework examples demonstrate the application of this approach using 6 STR loci and a sex discrimination locus. Although the lack of discrete allele identification could present difficulties in database comparisons between laboratories, the Kolmogorov-Smirnov test can be used to compare population samples. A computer program "STRquest" that uses the existing RCMP "Combien" foundation (G.R. Carmody) for rapid determination of floating bin allele frequencies will be demonstrated. The effect of FST and confidence intervals on STR frequency calculations although minor, has been noted.
Go to proceedings home page