pedbuildr now uses mirai for parallel
likelihood computations in reconstruct(), giving
substantial speedups in many cases. Users can enable parallelisation by
starting workers with mirai::daemons() before calling
reconstruct(). See ?reconstruct for
details.
The numCores argument has been deprecated, as
parallelisation is now controlled through the mirai framework.
More robust likelihood calculations. In particular, due to improved loop breaking in pedtools and pedprobr, many complex looped pedigrees that previously failed are now handled correctly.
buildPeds() is significantly faster than before in
the default case with linearInb = 0, i.e., when mating
between lineally related individuals is disallowed.
New S3 class pedCollection with for output of
buildPeds(), with separate print(),
plot() and subsetting methods
buildPeds() has a new argument
maxInbreeding indicating the highest tolerated inbreeding
coefficient in the constructed pedigrees. The default value is set to
1/16 (as in children of first-cousin parents), which is appropriate for
typical applications in human pedigrees. Set to 0 to disallow inbreeding
completely, or to 1 to allow any inbreeding.
New dataset Tutankhamun based on the pedigree and
genotype data published by Hawass et al. Ancestry and pathology in
King Tutankhamun’s family. Jama (2010).
General overhaul of code and organisation; the main functions run significantly faster now.