Regression Models for Allele Sharing in Affected Relatives
ABSTRACT:
Modelling of variation in identical-by-descent (IBD) allele
sharing using covariates can be useful to increase power to detect linkage,
identify covariate-defined subgroups that are linked to particular marker
regions, and improve the design of subsequent studies to localize genes and
characterize their effects in combination with other genes and/or
environmental factors. In this talk, we highlight and explore issues that
arise in studies of families with affected relatives. We have extended the
linear and exponential linkage likelihoods described by Kong and Cox [1997]
to model variation in non-parametric linkage (NPL) scores among
covariate-defined groups of families with affected relatives. We proposed a
likelihood ratio (LR) test for the covariate contribution and compared its
performance to a simple t-statistic for group mean NPL differences in the
case of a binary covariate. In simulation studies of locus heterogeneity in
families with affected siblings or affected cousins, we show how the
distribution of LR test statistics depends on the extent of linkage,
particularly in the presence of constraints on the parameters. On the other
hand, the distribution of the t-statistics may be biased by differences
between groups in information content. We recommend that constraints on the
parameters be applied with caution, and interpretation of covariate effects
in IBD allele-sharing models made with care.
About the Speaker
Shelley Bull is Senior Scientist in the Samuel Lunenfeld
Research Institute of Mount Sinai Hospital, Professor in the Department of
Public Health Sciences at the University of Toronto, and a Senior
Investigator of the Canadian Institutes of Health Research. Her research
program encompasses independent research in biostatistics and collaborative
research with investigators in epidemiology and molecular biology. Her
undergraduate and master's studies were in mathematics and statistics at
the University of Waterloo, with doctoral and post-doctoral training in
biostatistics and epidemiology at the University of Western Ontario. She is
currently project leader of the MITACS project in "Statistical genetic
modelling and analysis of complex traits".
References
Some suggestions for background reading are:
Kong and Cox (1997) Allele-sharing models: LOD scores and accurate linkage
tests, American Journal of Human Genetcs, 61, 1179-1188.
Mirea, Briollais, Bull (2004) Tests for covariate-associated heterogeneity
in IBD allele sharing of affected relatives, Genetic Epidemiology, 26, 44-60.
Olson, Witte, Elston (1999) Tutorial in Biostatistics: Genetic mapping of
complex traits, Statistics in Medicine, 18, 2961-2981.
