Skip to main content

Research

Research projects of the Igoshin Lab: microbial self-organization, bacterial stress response, cellular differentiation, biochemical information processing, viral cell-fate decisions, and PK/PD modeling.

Our research spans several interdisciplinary projects in computational systems biology. We use nonlinear dynamics, biophysics, data science, and bioinformatics to understand the behavior of biochemical networks and the principles of multicellular self-organization, and we work closely with experimental labs to test our models and theories.

Self-organization Stress response Differentiation Information-processing trade-offs Viral cell-fate decisions PK/PD modeling

Myxococcus xanthus aggregation and modeling

Mechanisms of microbial self-organization

How cells self-organize into complex tissues is one of the great puzzles of developmental biology, and a hallmark example of emergent behavior: complex patterns arising from simpler interacting components. Even the best-studied model systems still lack an understanding of the emergent properties that bridge developmental phenomena from molecules to cells, tissues, and whole organisms.

Biofilms formed by the soil bacterium Myxococcus xanthus are an excellent model system for emergent behavior. Under starvation, an M. xanthus biofilm starts a developmental program in which cells aggregate into mounds and then differentiate into distinct cell types. Complementing experimental research, we use computational biophysics and engineering approaches to build mathematical models of these phenomena, and we use recent advances in machine learning to connect the collective behavior of cells in a biofilm to their underlying genetic networks.

Selected papers

Architectures of bacterial stress-response networks

Evolutionary design principles of bacterial stress response

When exposed to harsh environments, antibiotics, or the immune response, bacteria survive by switching their gene-expression program into a “stress-response” mode. We want to understand how the biochemical networks that sense and respond to stress are organized in specific model systems, and to formulate general principles of stress response across bacterial species. Although their regulatory core is conserved, these networks vary in architecture. We look for the evolutionary design principles behind them by correlating these variations with different regulatory demands, focusing on bacterial sigma factors and two-component systems, two widespread mechanisms of master-level regulation of bacterial gene expression.

The stress-response network of Mycobacterium tuberculosis is critical for the survival of this intracellular pathogen within the host. To survive infection, M. tuberculosis must sense the stress conditions generated by the host immune system and adapt by reprogramming its gene expression and metabolism. With experimental microbiologists, we study how combinations of transcriptional and post-translational interactions shape the dynamics of stress-response networks.

Selected papers

Bacillus subtilis differentiation decisions

Functional requirements of cellular differentiation

Executing a differentiation program requires cells to weigh many internal and external cues and respond with the appropriate gene-expression program. The soil bacterium Bacillus subtilis has long been a model for prokaryotic differentiation. We use mathematical modeling to gain a systems-level understanding of how B. subtilis cells sense starvation and decide whether to start a unicellular (sporulation) or multicellular (biofilm) differentiation program.

We investigate how internal and external cues are dynamically encoded in the activity of master regulators of differentiation, and how different dynamics of a single master transcriptional regulator — that is, how its activity changes over time — can direct cells to alternative fates.

Selected papers

Kinetic networks illustrating speed–accuracy–dissipation trade-offs

Trade-offs in non-equilibrium biochemical information processing

The ability of enzymes and biochemical networks to read and propagate molecular information is key to the complexity of life. Yet thermal fluctuations and fluctuations in molecular copy numbers place fundamental limits on the precision of information processing. We use kinetic and statistical-physics approaches to understand how enzymatic mechanisms and free-energy landscapes evolved to balance speed, accuracy, and energy dissipation.

We apply these methods to DNA replication, transcription, and translation, as well as to gene regulation and chromatin dynamics. Our work on transcriptional fidelity shows how sequence-dependent kinetics, proofreading, and pausing set transcriptional error and speed. For the nuclease CRISPR-Cas9, we built a kinetic model of substrate selection, determined its parameters from single-molecule and high-throughput data, and used residence-time distributions to match measured conformational dynamics. This revealed intrinsic trade-offs between on-target cleavage speed and off-target specificity that can inform Cas9 engineering. We also study how chemical inhibitors and modulators change enzyme specificity, and we use data-driven parameter inference to connect theory with experiments. Together, these studies reveal fundamental constraints on biochemical information processing, with implications for synthetic biology and disease modeling.

Selected papers

Bacteriophage lambda infection and cell-fate decision

Viral cell-fate decisions: the biophysics of cellular counting

In many organisms, cells integrate information to select and transition to new phenotypic states, a process called a cell-fate decision. A classic model system for such decisions is infection of E. coli by bacteriophage lambda. A central question in the lambda decision is the role of viral copy number, which drives the outcome both as an initial condition and as a time-dependent quantity that changes as the virus replicates.

In collaboration with Prof. Ido Golding’s lab at the University of Illinois Urbana-Champaign, we develop physical models of how infected cells “count”, and of other open questions such as the role of stochasticity in the decision outcome.

Selected papers

Pharmacokinetic/pharmacodynamic model schematic

Mathematical modeling of pharmacokinetics and pharmacodynamics

Rational design of drug delivery requires precise knowledge of how a drug distributes through the body (pharmacokinetics, PK) and how it affects its target tissue (pharmacodynamics, PD). For protein therapeutics such as monoclonal antibodies and recombinant cytokines, these relationships are further complicated because a drug’s PD can affect its PK. And while preclinical animal models provide crucial insights, predicting how PK/PD data translate to humans remains a major challenge.

To address these complexities and improve predictive accuracy, we develop mathematical models built on data collected by our collaborators. We apply these models to understand and predict human treatment outcomes in promising areas such as immunotherapy, molecular glues, and GLP-1 receptor agonists.

Selected papers