Software version 3.31 (4/29/98)
Overview
Development of computer software for statistical genetic analysis can
be facilitated by the availability of software tools that can be used to
(1) verify the algorithms underlying a particular method of analysis (i.e.
statistical test), (2) determine empirical type I error rates for a statistical
test, (3) determine the power of a test, and (4) determine the robustness
of a test with respect to failures of underlying assumptions. The Genometric
Analysis Simulation Program (G.A.S.P.) is a software tool that can generate
samples of family data based on user specified genetic models. Data generated
can be as simple as a single sample of random individuals with a single
normally distributed trait or as complex as thousands of samples of extended
families with multiple traits based on a linear combination of major locus,
polygenic, common sibship environment and covariate components. Traits
can be generated based on a number of user specified components, and components
can be unique to a single trait or shared by multiple traits. The user
first specifies a list of all desired components and then creates each
trait by specifying the desired component weighted by its contribution
to the phenotypic variance.
G.A.S.P. can be used in two ways. First, data can be generated in a
standalone fashion. The resulting family data can be saved and then used
as sample data for demonstrating applications and methods of genetic analysis
or for testing and verifying newly developed algorithms in statistical
genetics. A simple driver ("dataonly") is provided for this application.
Second, data can be generated and analyzed immediately using an existing
statistical package. A driver can be designed to call subroutine versions
of widely available genetic analysis programs.
What can G.A.S.P. be used for?
- Verify analysis algorithms with respect to the underlying
theory
- Test the statistical validity of newly developed methods of genetic
segregation and linkage analysis and investigate the statistical properties
of the test statistics
- Determine the power and robustness of these methods
- Apply insights gained from these simulation experiments to ongoing
collaborative genetic analyses
|