[timothymillar]

Breaking Change:

• Changed the default prior used in assemble to a flat prior across genotypes.
  This is a more sensible prior in almost all situations because the previous
  Dirichlet-multinomial prior had to assume that all possible haplotypes (i.e.,
  combinations of SNVS) where equally likely to be present in the sample population.
  In general, this inflated the prior probability of heterozygous genotypes.
  Note that this is not a significant issue when using the call program with
  known haplotypes (and a reasonable estimate of their frequencies).
  The previous Dirichlet-multinomial prior can still be used via the
  --use-dirmul-prior argument. Consider defaulting flat prior across genotypes #139
• Changed the default prior used in call and call-exact to a flat prior across
  genotypes. This is a better default option than the previous Dirichlet-multinomial
  prior when sample inbreeding and prior allele frequencies are not specified.
  The previous Dirichlet-multinomial prior can still be used via the --use-dirmul-prior
  argument which is now used to set both the sample inbreeding and prior on allele
  frequencies. This new argument replaces the --inbreeding and --prior-frequencies
  arguments. Consider defaulting flat prior across genotypes #139