Modern biology is to increasing degree dependent on so called high throughput techniques, i.e. massively parallel experiments that generate a large set of readouts. Examples of such techniques are shotgun proteomics, yeast two hybrid, micro-arrays and next generation sequencing. A common challenge for these kinds of experiments is that the interpretation of the outcomes, as the individual measurements are of varying quality. I am aiming at increasing the yield and facilitating the interpretation of high-throughput experiments by using different machine learning methods such as support vector machines and dynamical Bayesian networks.