An implementation of a Machine Learning Framework for prediction of new drugs Side Effects. Firstly drugs are clustered with respect to their features description and secondly predictions are made, according to Bayesian scores. Moreover it can perform protein enrichment considering the proteins clustered together in the first step of the algorithm. This last tool is of extreme interest for biologist and drug discovery purposes, given the fact that it can be used either as a validation of the clusters obtained, as well as for the possible discovery of new interactions between certain side effects and non targeted pathways. Clustering of the drugs in the feature space can be done using K-Means, PAM or K-Seeds (a novel clustering algorithm proposed by the author).