image description


MAPGen Research Centers


The Knowledge Base and Coordination Center (USC)

The Knowledge Base and Coordination Center possesses strong expertise in bioinformatics, statistics, computer science, as well as clinical and biological expertise. The center has three major functions: (1) Develop a knowledge base on interconnections among diseases. We will systematically identify, integrate, and analyze the vast amount of public data to comprehensively describe the shared molecular mechanisms among diseases. We will establish a multi-dimensional disease connectivity map that can be interactively accessed via web interface. Using this knowledge base, we will perform a series of research projects on disease relationships and mechanisms.  (2) Develop a bioinformatics infrastructure for the MAPGen consortium. We will perform integrative analysis of data generated by different RCs as well those from public domains, in order to gain deep insights and fundamental understandings of the shared molecular mechanisms among the HLBS diseases. (3) Establish an Administrative Center to coordinate activities across RCs. We aim to synergize the effort across all RCs to achieve the goal of understanding the genetic mechanisms responsible for the interconnections among cross-organ diseases. A database server (DiseaseConnect, has been developed to facilitate the analysis and visualization of a comprehensive knowledge on mechanism-based disease connectivity.

Team: Jasmine Zhou (PI), Edward Crandall (PI),  Fengzhu Sun (Co-I),  Michael Waterman (Co-I), Andrey Rzhetsky (Co-I, U. Chicago), Preet Chaudhary (Co-I), Wenyuan Li (Scientific Coordinator), Jim Liu (Project Leader)


Harvard Center:

Since the 19th century, human diseases have largely been defined by the organ system in which they are most obviously manifest, and often so at end-stage.  The biomedical community now recognizes that many different diseases have common mechanisms and common endophenotypes or intermediate pathophenotypes (e.g., inflammation, thrombosis, apoptosis, fibrosis).  Based on this perspective of disease pathogenesis, the site of disease expression may be viewed a consequence of the local environment and of the differential expression of determinants of the intermediate pathophenotype in that environment.  We, therefore, propose as a central hypothesis that different complex diseases are governed by common network-associated determinants of common intermediate pathophenotypes, and that what differentiates these complex diseases from one another is the balance among the intermediate pathophenotypes, and the molecular context within which they are expressed.  To test this hypothesis, we will focus on three different diseases--acute myocardial infarction, venous thromboembolism, and acute ischemic stroke—and two intermediate pathophenotypes—inflammation and thrombosis—via three interdisciplinary specific aims.  First, we will develop network models of pathways that govern inflammation and thrombosis.  Concomitantly, we will utilize two large population-based whole genome scans to perform structured genetic analysis to identify components of inflammatory and thrombotic pathways related to the different diseases.  By combining this genetic analysis with network models, we will begin to construct subnetwork maps of elements of the ‘inflammasome’ and ‘thrombosome’ common to these diseases and elements that distinguish them from one another.  Second, using data sets derived from trials of the anti-inflammatory agent, rosuvastatin, and the antithrombotic agent, aspirin, in initially healthy individuals, we will examine the effect of therapeutic perturbation of the inflammasome and thrombosome on the incidence of each disease as determined by gene status.  We will also utilize key molecular mediators of the inflammasome and thrombosome common to and distinctive for these three diseases in correlative, iterative mechanism studies using relevant cell systems and animal models.  Third, we will integrate the network models of inflammation and thrombosis to develop predictive, probabilistic, multivariate models of manifestations of these diseases.

Team:  Joseph Loscalzo (PI), Albert-Laszlo Barabasi (Co-I), Dan Chasman (Co-I), Zak Kohane (Co-I), Paul Ridker (Co-I)


Stanford Center:

Overweight/obese Individuals are at increased risk of being insulin resistant (IR), and more likely to develop cardiovascular disease (CVD), type 2 diabetes (2DM), and obstructive sleep apnea (OSA). CVD and 2DM occur commonly in patients with OSA, leading to the view that CVD and 2DM are secondary to OSA. Alternatively, there is evidence that insulin resistance can lead to the development of OSA, similar to the pathogenesis of 2DM and CVD. Our research aim is to evaluate the possibility that insulin resistance not only contributes to the etiology of OSA, but is the common feature explaining why OSA, 2DM, and CVD form a clinical cluster. We also postulate that administration of pioglitazone (PIO), an insulin-sensitizer, to IR subjects with OSA will enhance insulin sensitivity, associated with clinical improvement and decreases in cardio-metabolic risk. Our study has three primary goals. 1) A comparison of specific measurements of insulin action, insulin secretion, and multiple cardio-metabolic risk factors in overweight patients with OSA with a weight-matched control group; 2) administration of PIO, or placebo, to IR patients with OSA, with a comparison of the two interventions on insulin sensitivity and secretion, clinical status of OSA, and changes in   in cardio-metabolic risk. 3) Comparison  of the same experimental end-points folowioing the addition of continuous positive airway pressure (CPAP) to PIO-treated and placebo--treated patients with OSA. A secondary goal is to evaluate the impact that changes in the activity of apelin, the endogenous ligand for the APJ receptor, have on the mechanistic link between excess adiposity and insulin resistance by comparing  plasma apelin levels in patients with OSA vs. the control group, as well as after the two interventions in IR patients with OSA. Adipose tissue biopsies will also be obtained in IR with OSA before and after each of the two interventions to evaluate changes in apelin modulation of glucose and lipid metabolism at the tissue level.

Team: Gerald Reaven, M.D. (PI), Clete Kushida, M.D. (Co-I),  Philip Tsao, Ph.D. (Co-I),  Sun Kim, M.D. Alice Liu, M.D.


Pennsylvania Center:

The University of Pennsylvania (UPenn) Center proposal utilizes ABO as a model glycotransferase system to define disease and cell-specific glycoproteomes as novel risk predictors for clinically important heart, lung, blood and sleep (HLBS) phenotypes. It also initiates the development of mechanism-based glycomic prediction and classification across HLBS disorders characterized by endothelial dysfunction, inflammation and thrombosis.  A striking finding of recent genome wide association studies (GWAS) is the reproducible and diverse association of the ABO glycotransferase locus with heart, lung, blood and sleep (HLBS) phenotypes. These include acute myocardial infarction (AMI), coronary artery disease (CAD), venous thrombo-embolism and LDL cholesterol as well as endothelial and thrombotic biomarkers such as circulating levels of VonWillebrand factor (VWF), Factor VIII, and E-selectin. Furthermore, we have extended these findings to show nominal associations of the ABO locus with acute lung injury (ALI). Taken together, these data suggest an important impact of ABO glycotransferase activity and its glycoproteomic modifications across diverse HLBS diseases. They also underscore an underappreciated link between complex carbohydrate modifications and a network of related HLBS disorders characterized by endothelial dysfunction, inflammation and thrombosis. Therefore, to define disease and cell-specific glycoproteomes, we will work with our collaborators at the Complex Carbohydrate Research Center (CCRC) at the University of Georgia to apply unbiased mass-spectrometry (MS) glycomic approaches and ABO as a model glycotransferase system. We will define, disease (AMI and ALI) and cell (platelets and endothelium) specific ABO glycoproteomes in order to develop glycopeptide markers of HLBS disease risk and cross-organ, mechanism-based phenotypes in HLBS. These glycoproteomes will be developed as novel risk predictors of clinically important HLBS traits, including incident AMI and ALI, and ultimately in glycomic reclassification of HLBS diseases.

Team: Muredach Reilly (PI, UPenn), Don Siegel (UPenn), Jason Christie (UPenn) , Parastoo Azadi (CCRC), Lance Wells (CCRC), Michael Tiemeyer (CCRC)


Pittsburgh Center:

Lung, heart, and vascular diseases are common causes of death that frequently occur in the same patients. They are usually diagnosed and treated as distinct entities, but may share common molecular mechanisms that respond to the same treatments. The objective of the Pittsburgh team is to use high throughput approaches and the extensive resources of well-characterized tissues in the University of Pittsburgh Cardiovascular Institute, the Simmons Center for Interstitial Lung Diseases and the Vascular Medicine Institute to identify new molecular phenotypes across and within disease  and organ boundaries. To identify such phenotypes we will:

  1. Identify tissue molecular signatures by analyzing failing and non-failing human RV tissue in IPAH or SPAH; failing and non-failing human LV tissue; human IPF and control lungs; lungs and pulmonary vessels of patients with IPAH and SPAH; and lung, RV, LV and pulmonary vessels from the same  patients from our unique warm autopsy program. We will perform mRNA and  microRNA expression profiling, validate key patterns and pathways by high throughput qRT PCR and generate cross organ tissue microarrays to perform high-throughput tissue protein validation and localization.
  2. Identify biomarkers of disease presence, stage and outcome, within and across organ and disease boundaries, in easily accessible peripheral blood by analyzing peripheral blood from patients with LV and RV failure, IPAH and IPF. We will perform mRNA and microRNA expression profiles and determine expression patterns that predict disease presence, state and outcome within and across organ and disease boundaries. 3. Generate a disease and mechanism relevant transcriptional map in RV and LV failure, IPAH and IPF by performing an integrated analysis of mRNA and microRNA expression patterns as well as clinical data with the use of advanced computational approaches, followed by cell culture and animal model validations of analytic predictions. RELEVANCE (See instructions): Lung, heart, and vascular diseases are common causes of death that frequently occur in the same patients. They are usually diagnosed and treated as distinct entities, but may share common molecular mechanisms that respond to the same treatments. We will identify these common mechanisms by analyzing patterns of gene expression in different diseases and organs, using advanced molecular and computational techniques.


Yale Center:

The glycosyl hydrolase 18 (GH18) gene family contains evolutionarily conserved chitinase-like proteins (CLP) that are presumed to play essential roles in biology. However, their functional profiles have just begun to be investigated. We recently demonstrated that the prototypic CLP (YKL-40 in man and BRP-39 in the mouse) is a circulating regulator of apoptosis, alternative (M2) macrophage activation and TGF-bi elaboration and tissue fibrosis. We also identified the first receptor for any GH18 moiety, IL-13 receptor(R)a2. Studies of patients with idiopathic pulmonary fibrosis (IPF) revealed elevated levels of plasma YKL-40 that correlated with adverse outcomes. After kidney transplant, urinary YKL-40 levels were increased in patients with delayed graft function (DGF). These findings led us to hypothesize that (a) elevated levels of YKL-40 are biomarkers of risk-stratification in IPF and DGF; (b) YKL-40 is produced by epithelial cells and macrophages and mediates its effects by stimulating TGF-b1 and M2 macrophage activation via IL-13Ra2; (c) YKL-40 and IL-13Ra2 are therapeutic targets in IPF and DGF. To test this hypothesis we will evaluate YKL-40 as a biomarker in two separate cohorts of patients, one with IPF and one with DGF. We will also evaluate the mechanisms it uses to mediate its cellular effects and the utility of YKL-40/BRP-39 and IL-13Ra2 as therapeutic targets in these disorders. We will; Aim 1. Define the levels of circulating YKL-40 and their roles as biomarkers in IPF. Aim 2. Define the levels of circulating & urinary YKL-40 and their roles as biomarkers in DGF. Aim 3. Characterize the location and molecular profile of YKL-40 in IPF and DGF. Aim 4. Characterize the relationships between YKL-40, IL-13Ra2 and TGF-|31 and the ability of YKL-40 to regulate monocyte lineage cell fate and accumulation in IPF and DGF. Aim 5. Characterize the roles of BRP-39 and IL-13Ra2 in murine lung fibrosis and ischemia/reperfusion kidney injury models. RELEVANCE (See instructions): We have identified a relationship between the chitinase-like protein YKL-40 and two different forms of maladaptive repair in humans. Idiopathic Pulmonary Fibrosis and Delayed Graft Function following renal transplantation. We will explore the utility of YKL-40 to serve as a predictor of progression in these diseases, define the mechanism through which YKL-40 exerts its effects on fibrosis, and determine whether this pathway is a therapeutic target in one or both of these diseases.