Our research in interdisciplinary computer science and informatics has gradually evolved from the computational analysis of biological networks and functional genomics to translational bioinformatics and systems pharmacology. For translational bioinformatics, we refer to the development of new techniques and databases that can improve the clinical diagnosis and treatment of complex diseases including cancers. For systems pharmacology, we refer to the development of new computational systems biology techniques and models that change how future drugs are developed, which has been known for an otherwise serendipitous, long, and expensive process.

While translational bioinformatics focuses on creating new biomedical knowledge to generate clinical impacts, systems pharmacology focuses on preclinical mechanistic modeling related to drug development. Both research areas converge on the computational representation, integration, and modeling of biological networks, which serve as the primary foundation for modeling complex bio-molecular systems. Both research areas may translate into significant economic development opportunities. By developing computational techniques and resources in these research areas in collaboration with experimental biologists, clinicians, and drug developers, our lab is focused on addressing the following related interdisciplinary bioinformatics research questions:

Can we use biological networks to organize and integrate genomics, functional genomics, and proteomics data, leading to subsequent extraction of biologically relevant information, e.g., disease molecular mechanism?

What biomolecular network characteristics are correlated with the onset and progression of complex biological diseases such as cancers, diabetes, and neurodegenerative diseases?

What advantages does the use of network modules have in performing disease biomarker discovery, compared with single biomarker finding or pure machine learning techniques?

Can we predict candidate drug’s efficacy, side effect, and the likely success of candidate drugs in clinical trials, using integrative network medicine and systems pharmacology approaches?

Under various disease progression, environmental stimuli, or drug perturbation conditions, how do we track underlying molecular changes of the disease using complex molecular systems models and subsequently leading to discovery of proper drug targets for interventions?

In 2010,  Dr. Chen described a rationale (Naylor and Chen, 2010, in Personalized Med.) why biomedical research and biopharmaceutical development communities need to embrace advanced informatics innovations, particularly when applying systems biology and network medicine to problems such as biomarker discovery and drug repositioning. Despite progress in genomics and other -omics sciences, our ability to unravel human complexity remains quite primitive, often taking the form of statistical data interpretations followed by experimental validations. Without the guidance of domain knowledge models, these findings are often biased (platform-specific and cohort-specific) and inconsistent (with non-overlapping solutions or difficult to cross-validate), therefore difficult to be adopted by decision makers who must make “go” or “no-go” decisions, in which tens of millions of dollars and patient in trial's lives are at risk.


Automated model-driven knowledge discovery systems developed by the Big Data technology community in web searching has not transpired into the life sciences (Chen et al 2013, Briefings in Bioinformatics), primarily due to insufficient data available to build quantitative systems biology models. Therefore, a real challenge is, “can qualitative or semi-quantitative network models be made useful to translational biomedical scientists?” Our answer is “yes” when primary concerns are more about developing biomarkers and drugs that work in heterogeneous populations and less about molecular disease etiology.

Supported by Indiana Center for Systems Biology and Personalized Medicine founded in 2008, our lab has been developing community network medicine and systems pharmacology database resources. We updated Human Annotated and Predicted Protein Interaction (HAPPI) database initially published in (Chen et al 2009, in BMC Genomics) to compile one of the most comprehensive quality-controlled compendia of human protein interaction data source in HAPPI 2.0.