Statistical modeling of genetic effects on behavior and its variability

Univ Of North Carolina Chapel Hill, Chapel Hill NC

Investigators

Abstract

? DESCRIPTION (provided by applicant): Anxiety disorders are highly prevalent and impose a high burden of morbidity on the citizens of the United States. Advances in the clinical management of these disorders can only come from advances in the understanding of their pathophysiology. Valuable targets for detailed molecular and pharmacological study have been identi?ed by the collective application of a variety of mouse genetics studies. One important study design, inbred strain association mapping (ISAM), offers exceptional genetic resolution for the identi?cation of genetic factors of interest, but suffers from an abundance of false-positive associations that cloud the interpretation of its results. In pilot analysis of an ISAM conducted by Dr. Tarantino, a statistical property of the measured anxiety traits was noted that strongly violates an assumption inherent in all existing analytic methods for ISAM. This assumption is that the phenotypic variability among same-strain individuals is constant across all strains. The violation observed is striking heterogeneity in variance across strains. This violation of the model assumption has been shown through simulation to produce an abundance of false-positive associations; therefore it greatly diminishes the reliability of existing methods to identfy genetic variants that increase or decrease anxiety traits. Aim1: I pro- pose to develop and apply methodology for ISAM that accommodates heterogeneity in strain variance and therefore more reliably identi?es genetic variants that increase or decrease anxiety traits. The heterogeneity n strain variance is not only a nuisance to be accommodated; it also permits the study of a relatively novel class of genetic variant - those that increase or decrease the trait variability o an individual. These variability-controlling genetic variants represent a promising avenue for identi?cation of homeostatic control mechanisms, epistatic interactions, and gene-by-environment interactions. Aim2: I propose to develop and apply methodology for ISAM that identi?es genetic variants that in?uence the variability of anxiety traits. Mouse genetics researchers have long recognized the complementarity between the ISAM study design and the F2 intercross mapping (F2M) study design: ISAM provides high genetic precision but suffers from an abun- dance of false-positive associations, while F2M provides robust evidence against false associations but suffers from low genetic precision. To date, however, their analysis is combined only informally, in a manner that does not derive the maximum bene?t from these valuable datasets. Aim3: I propose to develop and apply methodology for principled joint analysis of ISAM and F2M on anxiety traits. These methodological advances are based on the heterogeneity of strain variance observed in anxiety traits and are therefore directly relevant to the study of anxiety disorders. However, I have found marked variance heterogeneity in more than half of all phenotypes deposited in the major repositories for experimental mouse genetics data. Thus, the methodological developments proposed here may in fact advance the study of a wide variety of medically-important traits.