Two important aspects of statistical genetics research are development of new methods and application of such methods to datasets. In many cases, statistical research focuses on development, and ignores delivery of the methods to the potential users. Recently the P1 and colleagues have proposed a conceptual basis for unifying """"""""model-based"""""""" and """"""""model-free"""""""" linkage and linkage disequilibrium (LD) analysis for genetic epidemiology studies of dichotomous traits. The approach has been used in practice by the PT, in numerous practical applications, but can be very cumbersome to apply using the existing software packages, making them inaccessible to the broader community of scientists who might wish to apply them in practice. To this end, as part of this application, it is proposed to develop and distribute software which can automatically perform these analyses in a simple and straightforward manner. By making powerful techniques accessible (at least in their most general form) to those without substantial statistical and computing training, an important service will be provided to the community of gene mappers, and those trying to detect linkage and/or LD in their own datasets. Furthermore, such methods can allow for post-hoc analysis of errors in genotyping, errors in genetic marker maps, and errors in the assumed genotype-phenotype relationship at the locus under study. Joint analysis of multiple genes will also be incorporated in the software, through the use of statistical models recently described by the PT. Additionally, software packages will be developed for simulation of more realistic datasets, allowing investigators to more accurately predict the chance for successful mapping within a specified study design. Software to simulate the evolution of gene sequences in populations will be made available, such that users can evaluate the effects of different assumptions about population parameters on allelic complexity, gene diversity, and levels of LD, and software will be developed for simulation and ascertainment of pedigrees from populations, conditional on models of joint genetic and environmental etiology of common traits (qualitative and quantitative). These programs will be distributed as part of the expanded ANALYZE software package to be developed in this proposed project.
Showing the most recent 10 out of 19 publications
