Title:
Nonparametric Linkage Analysis via Sequential Imputation
Abstract:
Multilocus calculations using all available information on all pedigree
members are important in linkage analysis. Exact calculation methods for
analysis are limited in either the number of loci or the number of
pedigree members they can handle. Alternatives to exact calculation are
Monte Carlo approaches, such as sequential imputation. Unlike exact
methods, sequential imputation can handle both a moderate number of loci
and a large number of pedigree members. Sequential imputation is an
application of importance sampling in which ordered genotypes are
sequentially imputed locus by locus and then the inheritance vectors are
imputed conditional on these genotypes. The resulting inheritance vectors
together with the importance sampling weights are used to derive
consistent estimators of the linkage statistics of interest. This
presentation will focus on the use of sequential imputation in calculating
nonparametric linkage statistics, such as S_pairs or S_all. Results from a
simulation study to illustrate the potential gain in power using larger
pedigrees will be discussed. The results from sequential imputation will
be compared with analyses performed by the popular analysis package
GENEHUNTER. The GENEHUNTER analysis needed to drop between 38% to 54% of
the pedigree members whereas sequential imputation was able to use all
pedigree members. The power gains from using all pedigree members were
substantial under two of the three genetic models studied.