DREAM challenges (http://dreamchallenges.org/) gather participants from a wide community of researchers to analyze data provided to help solve fundamental, cuttingedge questions about systems biology and translational medicine.
I will show how Python and scikitlearn can be used to process data, try out standard machine learning techniques, and develop ad hoc solutions to address such challenges. This talk will be illustrated by the example of two challenges in which I took part, both times in teams that placed second. The first of these challenges was in toxicogenetics (https://www.synapse.org/#!Synapse:syn1734172), aiming at predicting the toxicity of a chemical on a cell line. The second (https://www.synapse.org/#!Synapse:syn1761567) was aimed at predicting the response of patients to rhematoid arthritis treatment.