Predicting Causal Effects in High-Dimensionsal Settings

Understanding cause-effect relationships between variables is ofgreat interest in many fields of science. An ambitious but highly desirablegoal is to infer causal effects from observational data obtained byobserving a system of interest without subjecting it to interventions. Thiswould allow to circumvent severe experimental constraints or to substantiallylower experimental costs. Our main motivation to study this goal comes fromapplications in biology.

We present recent progress for prediction of causal effects with directimplications on designing new intervention experiments, particularly forhigh-dimensional, sparse settings with thousands of variables but based ononly a few dozens of observations. We highlight exciting possibilities andfundamental limitations. In view of the latter, statistical modeling shouldbe complemented with experimental validations: we discuss this in the contextof molecular biology for yeast (Saccharomyces Cerevisiae) and the modelplant Arabidopsis Thaliana.