Figure 1

The eScience-Bayes approach. The red blocks represent the two aspects of eScience discussed in the main text: (light red) structured, federated, stored data and information that is retrievable with machines, and (dark red) networked computational power. The blue block represents the Bayesian part of the work flow: data setup and specification of the statistical model. The green blocks represent that locally produced data and information can be used together with the data and prior information retrieved from the Internet. (A) Interoperable machine-to-machine interactions, e.g. Web services, are employed to search the Internet for available information and data related to the biological or biomedical problem under study. (B) The biological or biomedical system is described in a Bayesian statistical model. The information collected in (A) is used to derive prior distributions of the parameters in the model, which summarize our a priori knowledge about the system. The data retrieved in (A) combined with any locally available data is used to update the prior distributions to a posterior distribution via the likelihood function (which represents the contribution of the data to the posterior distribution). (C) Fitting all but the simplest Bayesian models requires numerical integration, which is computationally highly demanding. Deploying the numerical integration on a HPC facility enables fitting large models in a manageable time.