Center for Bioinformatics and Systems Biology

Dr. Zhou is leading the Center for Bioinformatics and Systems Biology (CBSB) at Wake Forest University School of Medicine. By collaborating with biologists, pathologists, radiologists, mathematicians, engineers, and clinical doctors, the utmost goal of our scientific research is to be a leader in the development and applications of novel systems bioinformatic methods and molecular diagnosis as well as the integration of them into clinics and disease monitoring.

The center has dry labs and wet lab. The dry labs are focusing on developing advanced computational bioinformatics tools, clinical and translational informatics, biomedical imaging informatics tools, and systematic modeling tools to answer various biological and medical questions such as for biomarker and drug-target discovery, drug signature discovery, drug resistance mechanism discovery, disease mechanism, personalized medicine and drug delivery studies to improve basic, clinical and translation research, and meanwhile provide Bioinformatics and data analysis support for other investigators. The wet lab - experimental systems biology focuses on understanding stem cell niche, immune cell-cell interaction system, data generation and validation. Our goal is to attract NIH, NSF, DoD, and other foundation grants, publish original discoveries and support other group’s research. The research areas of the Center are mainly classified into three categories: Bioinformatics, Clinical and Translational Informatics, Systems Biology.

 

Research Team:

The CBSB Center has over 28 research fellows including 1 Full Professor, 3 Research Assistant Professors, 3 instructors, 11 postdocs, 6 software engineers, 4 Ph.D. and M.S. students.

 

I. Bioinformatics:

We have been working in genomics, genome-wide association studies, mass spectrometry proteomics, and metabonomics technologies to identify candidate disease biomarkers for many years. Recently we focus more on developing a software package, Mapping in Genomics Data Integration, to find the relationship between various types of genomic data, and then identify signal pathways and pathway signatures to define diseases with complex phenotypes. Our genomics team is also extensively working on the tool development of next generation sequencing data analysis, and particularly working on functional prediction of LncRNAs. Our group is also extensively working on the tool development for modeling signaling pathway map based on Phosphoproteomics data (reverse phase protein array). This technology is the key for us to research the signal transduction, drug combination treatment, and drug resistance studies. We have various Bioinformatics projects related to the following topics

  • Genomics Data Analysis: Array data analysis, Mapping between DNA methylation, Mutation, Transcription, and post-transcription, network inference, LncRNA functional prediction
  • Proteomics/Metabolics Data Analysis: Phosphoproteomics data analysis, signaling pathway modeling. Projects include P38 Downstream Pathway Discovery; Drug Resistance Study using Drug Combination Treatment Strategies; We are investigating the synergies of drug combinations in silico by considering that the drug combinations, which could block the feedback loops, bypasses and cross-talks.
  • Cellular Imaging Analysis, single cell data analysis
  • Data Integration – NIH LINCS program. We focus on drug-induced network signature discovery by integrating L1000 gene array data, Kinome Scan data, cellular imaging data, Cue Signaling data, and so on.
  • Big Data Science – develop computational tools specific for solving heterogeneous problems across different diseases and different data types at different scales.
 

II. Clinical and Translational Informatics (related to Systems Medicine)

1. PCORI project description: The Scalable Collaborative Infrastructure for a Learning Health System (SCILHS) project led by Harvard is funded by the Patient-Centered Outcomes Research Institute (PCORI) to cohere a clinical data research network comprised of 10 healthcare centers throughout the nation, including Wake Forest Baptist Medical Center (Dr. Zhou is the site PI). This system will cover more than 8 million patients and enable clinician and patient participation in research. It will build, test, and deliver informatics tools and join geographically and ethnically diverse network partners with patients to address the most important questions in three designated disease areas: pulmonary arterial hypertension, osteoarthritis, and obesity. By implementing necessary informatics technology and using open-source platforms that convert the electronic health record into a research tool, like I2B2 (Informatics for Integrating Biology and the Bedside) and SHRINE (Shared Health Research Information Network), SCILHS allows the sharing, querying and using of biomedical data. Plus, innovative "apps" in this program can capacitate point-of-care functions such as consent, enrollment, randomization, and patient-reported outcomes. Thus, SCILHS will generate agendas, identify research cohort from diverse population, inform participants about research, enroll them in trials study the cohort with ongoing bidirectional communication, and return research results. This will create a virtuous cycle where shared information drives improved care and returns knowledge to participants, sustaining engagement and enabling all healthcare stakeholders to learn from every healthcare encounter.

2. Big Data Science for Personalized Medicine (BD4PM): The explosive growth of biomedical Big Data provides enormous opportunities to revolutionize current clinical practices and biomedical research if the accompanying challenges of heterogeneity in those data can be addressed with novel informatics technologies. Needs of knowledge in translational research and clinical practice and knowledge provided by data-driven informatics technologies are often inconsistent, and new strategies and pipelines are required to bridge this gap. To address these challenges, we propose the BD4PM system, a biomedical Big Data informatics platform that allows fast adaptation of novel datasets by systematic data harmonization and knowledge management mechanisms, and expedites personalized medicine by providing user-oriented toolkits established on a scalable translational knowledge library. Our working hypothesis is that the similarity of signatures shared by patients and cell lines can reveal the underlying mechanisms of drug responses, and thus can be used in integrated models to optimize personalized medicines. We seek to investigate and develop a complete set of tools and resources that will enable signatures to be most effectively extracted from big data and applied to personalized medicine. Specifically, we unite information fusion and knowledge maturation strategies to address the challenges of data heterogeneity so that large scale signature extraction becomes feasible. We then develop novel algorithms to extract structured signatures from the integrated data to enable knowledge discovery. Finally, we investigate signature-based approaches to integrate advanced multi-scale models for personalized medicine. Through these innovative strategies, we will provide the biomedical science community with a complete set of tools and resources that allow researchers to develop and grow biomedical signatures as critically important knowledge for significant discoveries and in the process bridge key gaps between the growing amount of biomedical Big Data and the needs of translational research.

 

3. Develop Integrated Knowledge Environment (IKE) and IKE Studio for Big Data Science. A major challenge in big data arises from data heterogeneity. Data and tools from different sources are of disparate natures. Unified locating, querying, and accessing different data sources to address specific biomedical questions require novel data science approaches as well as the implementation of common standards. We are endeavoring to address these challenges with novel strategies and comprehensive solutions by developing a new perturbation-based minimum information standards, tracking data origins with the open provenance model, establishing a data warehouse of data discovery index (DDI), and develop the IKE (Integrated Knowledge Environment) Studio to provide an integrated knowledge discovery environment and to facilitate translational research with a contemporaneous and continuous supply of data and knowledge extracted from both existing and as yet emerging biomedical big data. In the IKE system the description of each dataset, in terms of the annotation in biomedical terms and the metadata describing experimental contexts, will be indexed into DDI tables following a set of common minimum information standards for various data entities. Thus, a DDI (without accessing the data sources) warehouse will be established to manage the descriptions of all LINCS (perturbation data)-related data sources in a unified way.

 

4. Biomedical imaging Informatics. We have been worked in the drug screening for anti-mitotic cell cycle and cell migration drugs using automated cell based assay high-content microscopy screen; high-content, automated microscopy for cell-based assays and the RNAi genome-wide screening technologies; high-content neuron assay based screening for discovering inhibitors for AD disease; and two photon and confocal microscopy image analysis for Neuron and Motor Neuron (Axon) related studies. We made substantial and original contributions on five softwares, namely itNETZ, Cell-IA, DCellIQ, GCellIQ, NeuronIQ, NeuriteIQ, and AxonTracker which have been developed and released. We made substantial contributions to the Microscopic image analysis and got four NIH R01s with other colleagues from 2005 to 2008 in these projects. Our lab is currently working on 3D in-vivo imaging informatics tool development, and medical image informatics, particularly in dental surgical planning and Craniomaxillofacial (CMF) deformation prediction based biomechanical properties and clinical information.

5. Clinical Data Based Intelligent System for the Control of Influenza. The goal of this research is to develop software to simulate a real world in order to help scientists and policy makers understand the transmission and spread of influenza and make effective evaluations for influenza control and prevention strategies. To accomplish it, we are developing a system in which the framework is an integrated social network and the components are created by an Agent-Based Model (ABM). The network has embedded dynamically-interacting, rule-based agents and therefore not only consists of the entire social structure (workplace, school, daycare, and public areas) but also has intelligent personal characteristics. Under certain rules, the system can result in extremely complex social behaviors. Hence, we call it an intelligent system. To form such a system, specifically, we will 1) build an integrated network to describe the social structure and household composition, based on inference from the U.S. Census; 2) describe individuals’ personal characteristics and health conditions in an agent-based model by incorporating clinical medicine, demography, and cognitive psychology into agent behavior; and 3) use a deep learning algorithm, under the supervision of real clinical data, to learn the parameters of the network and the agent-based model to create an optimized intelligent system. Through this systematic approach, this research aims to extend the reach and impact of intelligent system models on the spread and transmission of influenza. Once the system is validated and proven effective, it will be used to create the transmission of influenza in silico, analyze the agent behavior under conditions of influenza spread, and simulate the therapeutic and non-therapeutic interventions to prevent or mitigate it.    

 

III. Systems Biology:

Similarly to traditional biology, Systems biology needs clear hypothesis. To investigate the hypothesis, Systems Biology approach is to integrate biological experiments and systems modeling at different scale (molecular, cellular tissue, organ and population levels) seamlessly to answer various biological and medical questions. We currently have the following projects ongoing:

 

(1) Bridge the Gap Between the Cancer Mechanism and Population. Population-scaled molecular and signaling mechanism is helpful to design specific treatment strategies to reduce cancer incidence and mortality. Epidemiological studies have shown that women with Rheumatoid Arthritis have a significantly reduced risk of breast cancer and colon cancer. However, little is known of how immunosurveillance mechanisms, that is, immune-malignant cell interactions, prevent the formation and emergence of breast cancer. In this project, we are developing a novel systems biology approach to mechanistically bridge auto- and anti-cancer immunity in the context of breast cancer incidence using novel computational strategies and systems modeling and unique animal models that connect biological and population scales.

 

(2) NIH MSM project - Model Cancer Stem Niche and Tumor Growth. In this study, our hypothesis is that a multi-scale positive feedback loop remodels the biomechanical properties of myeloma stem cell niches, which in turn provides a pro-oncogenetic microenvironment for sustainable myeloma development: MICs (myeloma initiating cells) over-secret SDF1, which activates the SDF1/CXCR4 signaling pathway of neighbor BMSCs (Bone Marrow Stroma Cells), leading to BMSC contraction in MIC niches and consequently; the stiffened MIC niches promote proliferation and survival of MICs, and contribute to the MM growth and drug responses.  We integrate experiments and modeling to study the stiffer property of MIC and BMSC induced by SDF1/CXCR4 using gene array and RRPA data at intracellular level, and the model cell-cell interaction between MIC-BMSC-PCs-MMs and the MIC lineage model at intracellular level and, modeling the myeloma growth in tissue level.

 

(3) Multiscale Modeling of Immune Systems. Androgen ablation (chemical or surgical castration) is a first-line therapy for advanced Prostate cancer (PCa). Although androgen ablation is effective in treating primary PCa, a significant number of patients develop incurable castration-resistant disease. To successfully translate immunotherapy from the bench to the bedside, active and/or adoptive immunotherapies should be combined with blockade of inhibitory pathways. To fully characterize the impact of castration on the immune system, we have been developing a unique multipronged systems biology approach, which combined the dynamic data from in vivo studies with a predictive multi-scale (molecular level, cellular level, and tumor tissue level) multi-compartment (prostate compartment, blood vessel compartment, prostate-draining lymph nodes [PDLN] and spleen compartment) ordinary differential equations (ODEs) model. Based on these results, we hypothesize that 1) the immune microenvironment after castration involves innate (tumor-associated neutrophils [TANs]) and adaptive (cytotoxic T lymphocytes [CTLs] and Tregs) immune cells and two key cytokines (CCL20 and IL-2) produced by these cells; 2) Tregs and TANs regulate each other, thereby creating a reciprocal inhibitory loop that adversely affect CTL function; and 3) the interventions targeting Treg, TAN, IL-2, or CCL20 have distinct effects that block an inhibitory pathway. Our goal is to develop a predictive multi-scale multi-compartment model of the immune system (called M5I system) for systematically understanding the immune stimulatory and immune inhibitory mechanisms occurring after androgen ablation therapy. This model will predict the optimal intervention to prevent immune inhibitory pathways, thereby paving the way to improved immunotherapy.

(4) Systems drug delivery: Predictive modeling of Pharmacokinetics (PK) and Pharmacodynamics (PD)

We are collaborating with some faculties at Radiology Department and Nanomedicine Department to establish some mathematical and biophysical models to model drug delivery in organ level and drug release in extracellular microenvironment level (PK), as well as model the cell growth and death (PD) in response to the different drug treatment conditions. Our ultimate goal is to optimize drug delivery systematically by integrating multiscale mathematical modeling with the data generated from multiscale imaging and biological experiments. We are studying PK/PD at Macroscale using PET/SPECT, Mesoscale using MRI, and Microscale using intra-vital microscopy (IVM), and responses to drug treatment at cellular level. Doxorubicin is used as the prototype drug due to the well documented PK/PD of this drug; while the polymerized liposomes are used as the prototype drug delivery platform as this theranostic system can be easily modified to carry different contrasts and therapeutic agents. In order to achieve the research goal, the following three distinct and correlated projects are being carried out: multiscale imaging of drug delivery in different conditions, mathematical modeling of pharmacokinetics, and modeling of drug treatment response at target site.

(5) Multiscale Modeling from Genotype to Medical Imaging Traits for Inflammation Study. We are trying to develop Agent-Based modeling to bridge the gaps from genomics, proteomics, intercellular cellular imaging, and tissue imaging traits. Tumor-induced inflammation is conducted through the combat between various pro-inflammation and anti-inflammation mediators. T cells play an important role in regulating the chronic inflammation progress. Depending on the surrounding microenvironment, a naive T0 cell can differentiate into either a Treg cell, which secrets the anti-inflammation cytokines, or inversely a Th17 cell which secrets the pro-inflammation cytokines. The right figure is a simplified T0 cell differentiation signal pathway. Two kinds of cytokines, IL-6 and TGF- β, bound to the receptors respectively, will cause intracellular messengers into nucleus. In this pathway, if the transcription factor RORγt expresses strongly, the T0 cell will differentiate into Th17 cell. Otherwise, it will differentiate into Treg cell. After the differentiation, Treg cells and Th17 cells will affect each other and change the microenvironment by secreting the cytokines and chemokines, which will ultimately change the inflammation progress and the differentiation of T0 cells, and eventually will cause tissue level phonotypical changes which are expected to be measured using molecular imaging technology.

 

(6) Systematic Bone and Soft tissue Regeneration: We are applying multi-scale modeling strategies to simulate and guide the bone regeneration design. The overall objective of this project is to develop dual delivery nanocarriers as injectable synthetic bone grafts in order to reduce the risk of infection and improve bone regeneration for the treatment of battlefield-related contaminated injuries. There are three scales in the model of bone regeneration within scaffold: molecular scale, cellular scale and bone tissue scale. BMP2 signaling pathway at molecular scale induces the differentiation of osteoblast cells which result I the proliferation of osteoblast cells in the bone tissue, these osteoblast cells migrate from the bone tissue to scaffold, and the accumulated osteoblast cells on the scaffold surface promote bone growth within scaffold; During scaffold resorption, BMP2 is released from the scaffold which diffuses in the bone tissue and further stimulate the BMP2 signaling pathway at molecular. Another process is bone remodeling due to stress-strain mechanics. The utmost goal of this project is to apply this technology to teeth regeneration and kidney repair. 

Last Updated: 06-16-2014
USNWR 2013-2014Magnet Hospital RecognitionConsumer Choice2014 Best DoctorsJoint Commission Report

Disclaimer: The information on this website is for general informational purposes only and SHOULD NOT be relied upon as a substitute for sound professional medical advice, evaluation or care from your physician or other qualified health care provider.