International Mammalian Genome Society

The 13th International Mouse Genome Conference
October 31-November 3, 1999

Table of Contents * Structure * Bioinformatics * Sequence * Mapping * New Tools * Gene Discovery * Developmental * Mutagenesis * Functional Genomics

E12 Detecting Epistasis for Quantitative Traits in Inbred Line Crosses

Gary A. Churchill, Staff Scientist, The Jackson Laboratory, Bar Harbor, ME 04609; joint work with Beverly Paigen, TJL

Epistasis is a type of genetic interaction in which the allelic state at one locus can mask (or uncover) the effects of alleles at another. In modern usage, the definition of epistasis has been expanded to encompass any form of statistical interaction. However, the classical concept of masking effects has significant interpretative value. Epistasis analysis can suggest hypotheses about the nature of the pathway or network through which multiple genetic loci affect a trait. These hypotheses are testable using congenic strain constructs that carry combinations of alleles at two or more loci in a controlled genetic background. Although it seems likely that epistasis is important in the genetic control of many traits, previous work on quantitative trait mapping has generally neglected the phenomenon. Analysis tools are limited and no general guidelines for assessing epistasis have been adopted by the scientific community. We have developed a method to conduct a simultaneous search for pairs of interacting loci and implemented it in a software prototype. The method of simultaneous search examines all pairs of marker loci for association with the trait in a 2-dimensional genome scan. Significance of pairwise effects is assessed by a combination of permutation testing, to establish genome wide significance, and post-hoc testing, to determine the type of interaction. We have used this approach to analyze blood pressure data from reciprocal backcrosses between mouse strains A/J and C57BL/6J (Sugiyama, et al., 1999). We identified 8 loci and 4 interactions including all of the significant and suggestive loci from the standard main effects genome scan. Correspondence between mouse and human hypertension QTL is noted. It is also encouraging that all of the interactions that were identified by statistical analysis are explainable by the functions of candidate genes: they are either in the same biochemical pathway, or their proteins could serve as receptor and ligand. These results indicate the potential power of pairwise genome scans and epistasis analysis as a method of identifying and characterizing quantitative trait loci.


Abstracts * Officers * Bylaws * Application Form * Meeting Calendar * Contact Information * Home * Resources * News and Views * Membership

Base url
Last modified: Saturday, November 3, 2012