map.genotypes: Maximum posterior genotypes of unknown contributors

Description

For each marker, a ranked list of configurations of genotypes for some or all unknown contributors is returned. The list contains all configurations with posterior probability higher than some specified pmin.

IMPORTANT: This is the DNAmixturesLite package, which is intended as a service to enable users to try DNAmixtures without purchasing a commercial licence for Hugin. When at all possible, we strongly recommend the use of DNAmixtures rather than this lite-version. See https://dnamixtures.r-forge.r-project.org/ for details on both packages.

While the lite-version seeks to provide the full functionality of DNAmixtures, note that computations are much less efficient and that there are some differences in available functionality. Be aware that the present documentation is copied from DNAmixtures and thus may not accurately describe the implementation of this lite-version.

Usage

map.genotypes(
  mixture,
  pmin,
  U = seq_along(mixture$U),
  markers = mixture$markers,
  type = c("seen", "all", "unseen")
)

Value

A list, which for each marker contains the maximum posterior configurations of allele counts (genotypes) above the specified probabilities pmin.

Arguments

mixture

a DNA mixture

pmin

A list of the minimum probability to consider for each marker.

U

Optionally the indices of the unknown contributors of interest, specified as an integer vector.

markers

Optionally, a subset of markers.

type

It may be of interest to consider only the prediction of alleles in some subset of alleles. We allow

"seen": Consider only alleles that are seen in at least one EPG

"all"

Consider the entire allelic range

"unseen"

Consider only alleles that are not seen in any EPG (possibly redundant)

Details

Note that an error occurs if there are no configurations with probability higher than pmin. In this case, try a smaller pmin.

The function makes use of map.configurations, which localises the configurations of highest high posterior probability by simulating from the Bayesian networks until enough (at least mass 1-pmin) of the state space has been explored -- the computation time is thus dependent on how flat the posterior is, and how small pmin is. The simulation is used only to locate the relevant configurations; the computed probabilities are exact.

Examples

Run this code

data(MC18, USCaucasian, SGMplusDyes)
mix <- DNAmixture(list(MC18), k = 3, K = "K1", C = list(50), database = USCaucasian)
p <- mixpar(rho = list(30), eta = list(30), xi = list(0.07),
            phi = list(c(K1 = 0.7, U1 =0.2, U2 = 0.1)))
## Inlude the peak height information
setPeakInfo(mix, p)
## Marginally best genotypes for contributor U1
mpU1 <- map.genotypes(mix, pmin = 0.01, U = 1, type = "seen", markers = "D16S539")
summary(mpU1)
# \donttest{
## Jointly best genotypes for all unknown contributors
mp <- map.genotypes(mix, pmin = 0.01, type = "seen")
summary(mp) ## Profiles as genotypes rather than allelecounts
# }