RESEARCH PROGRAM OVERVIEW

 

Biological systems are beautifully complex and their nonlinear dynamics mediate the difference between health or disease, treatment success or treatment failure. The Yang Lab seeks fundamental, first-principles understanding into the molecular mechanisms underlying the pathogenesis of and therapeutic efficacy for chronic and infectious diseases.

 

We leverage advances in systems biology and biomedical data science to drive our discovery of mechanistic insights. We employ quantitative, live-cell, dynamic, high-throughput, and multi-OMIC experimental approaches. We couple these to network modeling, machine learning and bioinformatic analyses. Our research focuses on tuberculosis and heart failure.

EXPLAINABLE MACHINE LEARNING

Yang JH, Cell 2019

Recent advances in high-throughput experimental technologies and data analyses have enabled the unprecedented observation, quantification and association of biological signals with cellular and clinical phenotypes. However, current methods for inferring biological mechanisms from large datasets frequently encode such information in experimentally inaccessible "black-boxes". We are developing integrated network modeling and machine learning-based approaches to rapidly reveal causal mechanisms underlying therapeutic efficacy and disease pathogenesis.

​

Relevant Publications:​

CONTEXT-DEPENDENCE OF ANTIBIOTIC EFFICACY

Antimicrobial resistance is a growing threat to global health. Despite our knowledge of the primary targets for conventional antibiotics, it remains unclear why antibiotic treatment can sometimes fail. We are combining high-throughput assays, OMICS-characterization, fluorescence microscopy and animal experiments with network modeling, machine learning and bioinformatic analyses to explore how bacterial metabolism contributes to the context-dependence of antibiotic efficacy.

​

Relevant Publications:

Yang JH*, Bhargava P*, Cell Host Microbe 2017

TUBERCULOSIS RELAPSE AND LATENCY

NIAID

Tuberculosis (TB) is the leading cause of death from a single infectious agent from across the world. Standard treatment for TB involves combination therapy with at least 4 antibiotics for a minimum of 6 months. Even so, TB infection relapse rates are ~5% and patients may harbor undetectable latent infections lasting years before relapse. We are applying our white-box machine learning methods towards understanding mechanisms underlying the efficacy of TB antibiotics. We are also developing data-driven approaches for inferring mechanisms underlying TB relapse and latency from human-derived biospecimens. We experimentally validate these mechanisms in a shared BSL-3 facility in the Center for Emerging and Re-Emerging Pathogens.

​

Relevant Publications:

IMMUNOLOGICAL DECISION MAKING

Human cells possess the incredible ability to integrate many simultaneous (and sometimes contradictory) signals to make complex decisions in selecting between different cellular behaviors. This is especially true for innate immune cells which must sense, compute and respond to extracellular signals to migrate to sites of infection or damage and perform specialized tasks. We previously made the surprising discovery that antibiotics can also non-specifically act on immune cells, perturbing their metabolism, and interfering with their ability to clear pathogens. We are combining dynamic live-cell imaging experiments, phenotypic assays with network modeling and machine learning to understand how innate immune cells process extracellular cues.

​

Relevant Publications:

Yang JH*, Bhargava P*, Cell Host Microbe 2017

PATHOPHYSIOLOGIC CARDIAC CELL SIGNALING

Yang JH, J Mol Cell Cardiol 2014

Cardiovascular diseases are the leading causes of death around the world. First-line therapeutics target the same cell signaling pathways as those used by the body to adapt to changes in oxygen demand in healthy individuals. We previously demonstrated how network topology and subcellular compartmentation both regulate β-adrenergic signaling responses in cardiac myocytes, and how these are able to select between healthy physiologic responses vs. pathophysiologic responses to receptor activation. We are combining dynamic live-cell imaging experiments, phenotypic assays with network modeling and machine learning to understand cell signaling dynamics in cardiac cells in the context of heart failure.

​

Relevant Publications:

COLLABORATORS

​

David Alland, Rutgers New Jersey Medical School

Caleb Bashor, Rice University

James Collins, Massachusetts Institute of Technology

Veronique Dartois, Hackensack Meridian Health

Cheryl Day, Emory University

Allison Lopatkin, Barnard College

Bernhard Palssön, University of California, San Diego

Jyothi Rengarajan, Emory University

Jeffrey Saucerman, University of Virginia

Padmini Salgame, Rutgers New Jersey Medical School

David Sherman, University of Washington

Graham Walker, Massachusetts Institute of Technology

Jin Zhang, University of California, San Diego

RESEARCH FUNDING

​

National Institutes of Health

Rutgers New Jersey Medical School