ldDesign: Design of experiments for detection of linkage disequilibrium

R package for design of experiments for design of genome-wide association studies. Version 2 incorporating quantitative traits and case-control studies. The Bayes factor should be chosen large enough to give respectable posterior odds. This requires Bayes factors of the order of 10^6 in genome-wide association studies where prior odds are low. Sample sizes needed to get this strength of evidence are substantially higher than those from traditional power calculations. The corresponding threshold for p-values is substantially lower than commonly used. For quantitative traits ldDesign uses an existing deterministic power calculation for detection of linkage disequilibrium between a bi-allelic QTL and a bi-allelic marker, together with the Spiegelhalter and Smith Bayes factor to generate designs with power to detect effects with a given Bayes factor. For case- control studies an asymptotic approximate Bayes factor is used to derive an analytical power calculation in dominant, recessive, additive and general genetic models.

Downloads: