A very active field of research in Bioinformatics is to integrate structure in Machine Learning methods. Methods recently developed claim that they allow simultaneously to link the computed model to the graphical structure of the data set and to select a handful of important features in the analysis. However, there is still no way to simulate data for which we can separate the three properties that such method claim to achieve. These properties are: (i) the sparsity of the solution, i.e., the fact the the model is based on a few features of the data; (ii) the structure of the model; (iii) the relation between the structure of the model and the graphical model behind the generation of the data.
