Descent GraphsinPedigree Analysis: Applications toHaplotyping, Location Scores, andMarker-Sharing Statistics

Summary Theintroduction ofstochastic methods inpedigree anal- ysis hasenabled geneticists totackle computations in- tractable bystandard deterministic methods. Until now these stochastic techniques haveworked byrunning a Markovchain onthesetofgenetic descent states ofa pedigree. Eachdescent state specifies thepaths ofgene flowinthepedigree andthefounder alleles dropped downeachpath. Thecurrent paper follows upona suggestion byElizabeth Thompson that genetic descent graphs offer amoreappropriate space forexecuting a Markovchain. A descent graph specifies thepaths of gene flow butnottheparticular founder alleles traveling downthepaths. This paper explores algorithms forim- plementing Thompson's suggestion forcodominant markers inthecontext ofautomatic haplotyping, esti- mating location scores, andcomputing gene-clustering statistics forrobust linkage analysis. Realistic numerical examples demonstrate thefeasibility ofthealgorithms.