FDC Statistical Analysis & Population Genetics 2 — Questions and Answers
Question 1: A Likelihood Ratio (LR) in forensic DNA statistics compares:
- The probability of the evidence given the prosecution hypothesis versus the probability of the evidence given the defense hypothesis (Correct answer)
- The probability of a match given innocence versus guilt
- The combined profile frequency across two separate loci
- The ratio of heterozygote to homozygote frequencies
Correct answer: The probability of the evidence given the prosecution hypothesis versus the probability of the evidence given the defense hypothesis
The LR is defined as P(evidence | prosecution hypothesis) / P(evidence | defense hypothesis), quantifying how much more likely the evidence is under one scenario compared to the other.
Question 2: When a DNA mixture is observed and the contributor number is uncertain, which statistical approach can account for all possible numbers of contributors?
- Probabilistic genotyping software using continuous models (Correct answer)
- The product rule applied to all alleles present
- The ceiling principle with average allele frequencies
- A simple RMP calculated from visible alleles only
Correct answer: Probabilistic genotyping software using continuous models
Probabilistic genotyping software using continuous models evaluates the data under all plausible contributor number hypotheses and weights them appropriately, providing more rigorous mixture statistics than binary methods.
Question 3: The Combined Probability of Inclusion (CPI) in mixture statistics represents:
- The probability that a randomly selected, unrelated person could be a contributor to the mixture (Correct answer)
- The probability that the known contributor is the source of the major component
- The likelihood ratio favoring inclusion of a specific individual
- The number of individuals whose profiles are consistent with the mixture
Correct answer: The probability that a randomly selected, unrelated person could be a contributor to the mixture
CPI is the product of single-locus Probabilities of Inclusion across all tested loci, representing the overall chance that a random, unrelated individual would not be excluded as a possible contributor to the mixture.
Question 4: Which of the following best describes the 'sampling uncertainty' concern in forensic DNA statistics?
- Allele frequencies estimated from finite reference databases may not perfectly represent the true population frequencies (Correct answer)
- The DNA sample collected at the crime scene may be too small to yield a full profile
- Random errors in the capillary electrophoresis instrument affect allele sizing
- Uncertainty about which suspect contributed to the mixed profile
Correct answer: Allele frequencies estimated from finite reference databases may not perfectly represent the true population frequencies
Sampling uncertainty arises because reference population databases contain a limited number of individuals, so estimated allele frequencies carry statistical uncertainty that can be addressed using confidence intervals or conservative estimates.
Question 5: If a forensic analyst calculates an RMP of 1 in 10 quadrillion for a 20-locus CODIS profile, this statistic means:
- The profile is expected to occur in approximately 1 out of every 10 quadrillion unrelated individuals (Correct answer)
- The probability that the suspect committed the crime is 10 quadrillion to one
- Laboratory error is 1 in 10 quadrillion
- The match probability exceeds the world population and is therefore meaningless
Correct answer: The profile is expected to occur in approximately 1 out of every 10 quadrillion unrelated individuals
RMP expresses profile rarity — 1 in 10 quadrillion means that among unrelated individuals, we expect approximately 1 in 10^16 to share that exact profile by chance, not that guilt is established.
Question 6: In forensic DNA statistics, the 'ceiling principle' proposed by NRC I involved:
- Using a minimum allele frequency of 5% (or 10%) regardless of the observed frequency to provide conservative estimates (Correct answer)
- Capping the RMP at 1 in 1 billion regardless of calculated values
- Limiting the number of loci used in calculations to avoid overclaiming
- Applying an upper confidence limit to the population database allele counts
Correct answer: Using a minimum allele frequency of 5% (or 10%) regardless of the observed frequency to provide conservative estimates
The ceiling principle proposed by NRC I set a minimum allele frequency floor (e.g., 5% or 10%) to guard against underestimates from small or unrepresentative population samples, though it was later superseded by the product rule with substructure corrections.
Question 7: Under the NRC II framework, which major population databases are typically used for U.S. forensic DNA statistics?
- Caucasian, African American, and Hispanic reference populations (Correct answer)
- Asian, European, and Middle Eastern populations only
- A single universal population database
- Individual state-level population databases
Correct answer: Caucasian, African American, and Hispanic reference populations
NRC II and standard U.S. forensic practice recommend calculating RMP across major population groups — typically Caucasian, African American, and Hispanic — and reporting the least favorable or all values to address population of origin uncertainty.
FDC Statistical Analysis & Population Genetics 2
A Likelihood Ratio (LR) in forensic DNA statistics compares: