Biophysical Modeling of Genome Dynamics and the Humoral Immune Response
Name
kannan-dkannan-phd-physics-2026-thesis.pdf
Size
25.3 MB
Format
Adobe PDF
Checksum (MD5)
9f0fa849845c97ed52550b587675fccb
Author(s)
Kannan, Deepti
Advisor(s)
Chakraborty, Arup K.
Kardar, Mehran
Date Issued
May 2026
Publisher
Massachusetts Institute of Technology
Abstract
Biological function emerges from complex dynamical processes across physical scales. The human body is composed of trillions of cells, each containing over two meters of DNA densely packed into a nucleus only ten micrometers wide. The structure and dynamics of the genome and its associated proteins (chromatin) control which genes are turned on or off, regulating cell identity and cell state. On larger scales, specialized immune cells interact and coordinate their activities in order to generate immune responses that protect the body from diverse pathogens. In both contexts, known microscopic interactions — between biomolecular species (DNA, RNA, proteins) within cells, or between immune cells within tissues — collectively give rise to a macroscopic response (cell state or immune protection). Approaches from statistical physics provide a natural framework for analyzing how these microscopic interactions produce emergent macroscopic responses. In this thesis, we develop statistical physics models of genome dynamics and the humoral immune response and make experimentally testable predictions. In the first set of projects, we investigate how active, energy-consuming processes influence the folding and dynamics of chromatin. Many chromatin-associated enzymes and motor proteins consume chemical fuel in the form of adenosine triphosphate (ATP), generating active kicks to the polymer backbone which can vary along the sequence. Using simulations and theoretical analysis, we show that patterns of correlated active forces can produce polymer folding patterns similar to those observed in experimental measurements of chromatin structure. These folding patterns are indistinguishable from predictions of purely equilibrium polymer models. We therefore identify dynamical signatures that can distinguish active from passive mechanisms of genome folding. We also demonstrate that condensates (phase-separated assemblies) of transcriptional proteins can show directed motion towards sites of active RNA transcription by sensing gradients of RNA molecules. Transcriptional condensates are often associated with enhancers, regulatory DNA elements that are genomically distant from the gene promoters they control. As a result, condensate directed motion towards promoters offers a physical mechanism of enhancer-promoter communication and subsequent gene regulation. In the final project, we develop a mechanistic model to investigate how broadly neutralizing antibody therapy of people with HIV could lead to post treatment viral control. Our results highlight how feedback from early antibody responses can shape the subsequent evolution of more diverse and therefore more protective antibody responses during treatment. This modeling framework reconciles diverse clinical and immunological observations and suggests potential strategies for future therapeutic interventions.
MIT Department
Massachusetts Institute of Technology. Department of Physics
Terms of Use
In Copyright - Educational Use Permitted
Copyright retained by author(s)
Persistent DSpace Link
