Name: Rogers

Firstname: James

Title: Statistician

Institution: Cereon Genomics

Street: 45 Sidney Street

City: Cambridge

Zip-Code: 02139

Country: USA

Phone: (617) 551-8073

Fax: (617) 551-1920

Email: james.a.rogers@cereon.com

Authors: James A. Rogers (joint work with Jason C. Hsu and Shili Lin)

Title: Applying the Partitioning Principle to Genome Scans

Abstract: Genome scans may now test hundreds of markers for linkage with a given trait. Under the traditional formulation of testing for linkage, exact multiplicity adjustment is difficult because of the complicated nature of dependencies. Aside from the multiplicity problem, traditional linkage analysis does not allow any confident inference as to the location of the polymorphism affecting the trait within chromosomes. We present a new approach to linkage analysis. Specifically, we use the partitioning principle (Stefansson, Kim, and Hsu, 1984; Finner and Strassberger, 2002) to construct an exact confidence set for the locus of the disease gene. The confidence set is constructed in such a way that the relevant multiplicity is far less than the number of markers that are tested. If the trait is known to be monogenic then the relevant multiplicity is one, i.e. no multiplicity adjustment is necessary.

We evaluate the performance of our confidence set in the context of simple data types.

References: Stefansson, Kim, and Hsu (1984). Statistical Decision Theory and Related Topics, Berger & Gupta eds., Springer-Verlag.

Finner and Strassberger (2002) Annals of Statistics Vol. 30, No. 4 August 2002