A Bayesian LOD Score for Linkage Analysis of Complex Diseases.pdfVIP

A Bayesian LOD Score for Linkage Analysis of Complex Diseases ∗ ∗ Ma’ayan Fishelson, Dmitry Rusakov, Dan Geiger Computer Science Department, Technion Haifa 32000, Israel January 14, 2004 Keywords: Complex Diseases, MMLS, Linkage ease genes and their interactions which govern what Analysis, Lod score, SUPERLINK. is termed complex diseases. One method that has been pursued in the genetics literature for analyzing Abstract complex diseases is to compute the likelihood of odds (LOD score) assuming two inheritance models of the Motivation: The emphasis of genetic linkage anal- disease: recessive or dominant, each with 50% pene- ysis is rapidly shifting from the task of locating trance, commit to the model that produces the higher genes responsible for mendelian diseases, for which maximum score, and correct for multiple tests by sub- one major gene causes the disease, to identifying tracting 0.3 from the maximum score (Hodge et al., a collection of genes and their interactions which 1997; Greenberg et al., 1998). This method, called govern what is termed complex diseases. MMLS-C, has been shown to work well in simulations Results: We propose a Bayesian averaging approach (Greenberg et al., 1998; Abreu et al., 1999; Abreu based on a known method called MMLS-C . The et al., 1999; Greenber