The Genometric Analysis Simulation Program (G.A.S.P.) is a software tool for generating samples of family data based on user-specified genetic models. Quantitative traits are defined in terms of a combination of user-defined components, including a major-locus (or loci), a polygenic component, common sibship environment, and/or other covariates. In addition, loci corresponding to discrete traits or markers can be generated. However, complications such as missing data, genotyping errors reduced penetrance, genetic heterogeneity, epistasis, and non-random ascertainment are not handled directly by this software. G.A.S.P. is run in combination with a driver that formats the generated data and can also be used to add layers of complexity to the model. Driver-based trait modifications include: functional transformation, classification of a quantitative trait according to a threshold, and randomly determined expression of a discrete trait (reduced penetrance). More complicated trait models allow for the contribution of one locus to depend on the genotype at another locus. Non-random ascertainment can be modeled by discarding families that do not meet the desired criterion. The driver can allow for mixtures of populations, modeling partial association or genetic heterogeneity, by maintaining separate model specifications and random number streams for each population and including families from different populations in the desired proportions. Most simulation studies require the use of thousands of samples from each population model considered. It is necessary to generate data, invoke analysis programs, extract relevant output, and accumulate results for each replicate. A unix shell script is used in combination with G.A.S.P. and appropriate driver programs to accomplish these tasks. G.A.S.P. is currently being used at over 60 sites worldwide.
