Profile Interpretation, Mixtures, and Statistical Calculations
Interpretation
Start with the number of contributors, then assess peak balance, allele sharing, and stochastic effects. A single-source profile should show no more than two alleles per locus and reasonable peak height ratios; tri-allelic patterns and imbalanced peaks suggest mixtures or artifacts. Drop-out and drop-in require a stochastic threshold. Traps: counting alleles does not determine contributor number; a major component can mask a minor; conditioning on a known contributor is not the same as assuming guilt.
Statistics
For single-source profiles, report a random match probability (RMP) or likelihood ratio (LR). For mixtures, use CPI, RMP, or probabilistic genotyping. LR equals the probability of the evidence given the prosecution hypothesis divided by the probability given the defense hypothesis. Population databases require Hardy-Weinberg and linkage equilibrium assumptions; apply a theta correction for subpopulations. Traps: do not multiply across loci if loci are linked or the profile is not independent; do not attach a statistic to an exclusion; CPI is not a likelihood ratio and cannot identify a contributor. Relatives reduce the strength of a match.
Worked example
Single locus, heterozygote 8,10; allele frequencies 0.1 and 0.05. RMP = 2(0.1)(0.05) = 0.01. If the alternative is a random unrelated person, LR = 1/0.01 = 100. If the alternative is a sibling, use conditional probabilities, and the LR drops substantially. Report the hypothesis pair explicitly, not just the number.