System-Subsystem Dependency Network for Integrating Multicomponent Data and Its Application to Health Sciences.

With the ever-increasing demand of collecting and analyzing a large volume of data collected from different sources, computational models and methods-ones that allow for the integration of data sets to help understand a phenomenon from a broad, comprehensive systems perspective-are called for. Two features are commonly observed in large and complex systems. First, a system is made up of multiple subsystems. Second, there exists fragmented data. The methodological challenge is to reconcile the potential parametric inconsistency across individually calibrated subsystems. This study aims to explore a novel approach, called system-subsystem dependency network, which is capable of integrating subsystems that might have been individually calibrated using separate data sets. From a health informatics perspective, the method can be seen as a way to integrate heterogeneous data sources, especially from relatively well-structured clinical study data. In this paper, we compare several techniques for solving the methodological challenge. Additionally, we use data from a large-scale epidemiologic study, as well as from two large clinical trials to illustrate the solution to the inconsistency of overlapping subsystems and the integration of data sets.