Skip to contents

Monte Carlo exclusion test for candidate source populations

Source: R/test_source_exclusion.R
test_source_exclusion.Rd

Test whether an individual's multilocus genotype is unusually improbable under each candidate source population. This complements the relative ranking returned by [assign_individuals()]: an individual can rank first for a source while still being incompatible with every sampled source.

Usage

test_source_exclusion(
  result,
  data = NULL,
  simulations = 10000L,
  alpha = 0.01,
  frequency.floor = NULL,
  random.seed = NULL,
  verbose = TRUE
)

Arguments

result

Result returned by [assign_individuals()] using the native called-genotype likelihood engine.

data

Optional genomic data used for the assignment. Supplying it lets simulations reproduce the exact non-missing marker set of each individual. Without it, the function matches only the individual's number of scored markers by sampling from markers available in the candidate source. Default: data = NULL.

simulations

Number of simulated genotypes per candidate source. Default: simulations = 10000.

alpha

Lower-tail probability used for exclusion. Default: alpha = 0.01.

frequency.floor

Minimum allele frequency used to avoid zero probabilities. With `NULL`, use `1 / (2 * N + 1)` from the reference count for each source-marker combination. Default: frequency.floor = NULL.

random.seed

Optional simulation seed. Default: random.seed = NULL.

verbose

Logical. Display progress messages. Default: verbose = TRUE.

Value

A list containing one row per individual and candidate source with the observed likelihood, Monte Carlo lower-tail probability, exclusion call, and simulation threshold; an individual summary; and simulation settings.

Details

For each source and marker, genotypes are generated under Hardy-Weinberg proportions using the reference alternate-allele frequency. The same likelihood model used for assignment scores each simulated genotype. The Monte Carlo probability is `(b + 1) / (B + 1)`, where `b` is the number of simulated likelihoods no greater than the observed likelihood and `B` is `simulations`.

This is a source-exclusion diagnostic, not proof of origin or migration. Results depend on representative reference sampling, allele-frequency estimates, Hardy-Weinberg assumptions, genotyping error, missingness, linkage disequilibrium, family structure, and whether the true source was sampled. Linked markers make the product likelihood overconfident. Repeat the test on linkage-pruned data or biologically defined marker blocks. Candidate inversion regions should be analysed explicitly because suppressed recombination can contribute many correlated markers.

The current simulator is for called diploid genotypes. GL/PL assignment needs coverage-aware read or genotype-likelihood simulation and is not silently approximated here.

References

Cornuet JM, Piry S, Luikart G, Estoup A, Solignac M (1999). New methods employing multilocus genotypes to select or exclude populations as origins of individuals. Genetics, 153, 1989-2000.

Paetkau D, Slade R, Burden M, Estoup A (2004). Genetic assignment methods for the direct, real-time estimation of migration rate: a simulation-based exploration of accuracy and power. Molecular Ecology, 13, 55-65. doi:10.1046/j.1365-294X.2004.02008.x .

Manel S, Gaggiotti OE, Waples RS (2005). Assignment methods: matching biological questions with appropriate techniques. Trends in Ecology & Evolution, 20, 136-142. doi:10.1016/j.tree.2004.12.004 .