Development of a Framework for Constructing a Virtual Physiological Human with the Integration of COntext Specific Directed Associations (CODA)

Abstract

Construction of a virtual physiological human system by reconstructing biological networks is proposed as an alternative to traditional drug development process. In silico methods using biological networks in human have been used for resolving current issues of traditional drug discovery like high costs and low success rates. However, most of previous approaches for reconstructing biological networks have considered only one relation type although there are various types of relations not only in molecular level but also within over-molecule levels such as biological processes and diseases. In addition, each tissue specific network in previous works is isolated and it is not appropriate to analyze diseases associated to multi-tissues like type 2 diabetes (T2D).

Here, we develop a framework for reconstructing genome-scale biological networks into a whole body level. We collected and integrated biological interactions in bio-synergy modeling language (BSML) format, which can contain anatomical context information and multi-components. Based on BSML format, we constructed context specific directed associations (CODA) repository comprising various types of interactions including not only molecular interactions such as tissue specific intracellular interactions and intercellular interactions but also associations between molecules and over-molecules. All of associations in CODA repository have anatomical contexts covering 71 tissue-cell types. To verify the possibility of uses of CODA to drug development, we selected T2D as a case study. We simulated effects of drugs on T2D with our CODA system, drug-drug target interaction information, and Petri Net.