The advantage of decoupling these two parts of the analysis in this manner is that each can be changed independently from the other, simplifying the development of changes and extensions to either part of the model. Moreover, since the second part only uses the output from the first part and not the raw genotype data they do not need to be performed at the same time or place, making it much more straightforward to perform multiple gene-set analyses on the same data or to analyse multiple data sets across a large-scale collaboration.
