The framework for MAGMA is built with future extensions in mind. Because of the two-tiered structure of the gene-set analysis, alternative gene analysis models are straightforward to implement and are automatically available for use in the gene-set analysis. Similarly, the linear regression structure used to implement the gene-set analysis offers a high degree of extensibility. At present it enables analysis of continuous gene-level covariates as well as conditional analysis of gene-sets correcting for possible confounders, and the analysis of the CD data demonstrates that correction for confounders such as gene size and gene density is indeed strongly advised. The model is easily generalized to much more general gene-level linear regression models to allow for simultaneous analysis of multiple covariates and gene-sets, opening up a wide range of new testable hypotheses.
