When disease alleles of monophyletic origin are sampled from a population,
haplotypic associations are observed at linked markers, due the coancestry
of the sample.  Give a set of realized coancestries, a Monte Carlo likelihood
for fine-scale disequilibrium mapping can be estimated.  However, ancestries
of a sample of a given disease allele are very different from population
coalescents -- selecting genes of given allelic type is a very strong form
of ascertainment bias.  I describe one method of realizing these ancestries,
conditional on current disease allele frequencies, under an assumption of
monophyletic origin at some specific past time.
  This is joint work with Jinko Graham, Dept Biostatistics, UW.