lrMix {relSim}

Calculate locuswise likelihood ratios for two person victim/suspect mixtures

Description

Calculates the likelihood ratio for pairs of profiles under the propositions H_p:\qquad V+S and H_d:\qquad V+U, where V, S and U are the victim, the suspect and someone unrelated to the suspect respectively. The calculation does not employ \theta so there are no assumptions about the subpopulations of the contributors.

Usage

lrMix(profiles, Freqs)

Arguments

profiles

A vector of profile lists, from randomProfilePairs. randomPCPairs and randomSibPairs also work but should not really be used as the calculations do not take account of the relationship between the two individuals.

Freqs

A list containing elements freqs, loci and counts. The element freqs is a list of vectors of allele frequencies at the loci listed in loci. These frequencies are used to evaluate the LR

Value

A matrix of LRs calculated at each locus for every pair of profiles. Note this is the set of N profile pairs supplied in profiles, not a pairwise comparison.

Author(s)

James M. Curran

Examples

data(USCaucs)
p = randomProfilePairs(USCaucs, 10000)
log.lrs = log10(lrMix(p, USCaucs))
boxplot(log.lrs, las = 2)