Statistical Modelling of Biological Networks

Опубликовано 7 сентября 2016, 16:19

New experimental techniques in molecular biology make it possible to probe cellular processes in unprecedented detail. In particular, information on the activity profiles of genes is available from large scale microarray or reporter assay experiments. To infer detailed models for cellular processes poses formidable inference problems, if models beyond simple clustering or association graphs are envisaged. I will discuss some ideas and examples for the inference of static as well as dynamic models for gene regulation using a Bayesian model selection framework. The fact that most data sets are high dimensional but comparatively small suggests exploiting the simplicity and flexibility of Gaussian processes for modelling nonlinear relationships in dynamical models.