(Mathematical modeling of mRNA m6A dynamics)Epitranscriptomics refers to modifications to RNA molecules, particularly mRNA, that influence gene expression at the post-transcriptional level. N6-methyladenosine (m6A) is the most prevalent internal modification in eukaryotic messenger RNA (mRNA) and regulates critical RNA activities including stability and translation. In cancer, it has become increasingly evident that m6A modifications significantly influence tumorigenesis, progression and metastasis by altering the expression of oncogenes and tumor suppressor genes. However, the exact mechanisms of m6A dynamic regulation are not fully understood. In this project we will develop mathematical models aimed at describing/exploring m6A dynamics and elucidating its role in cancer progression, using mechanistic biological principles as well as published data. Understanding the dynamics and role of m6A modifications in cancer will ultimately open new avenues for developing therapeutic targets and biomarkers for prognosis prediction.

Cell fate determination is influenced by a variety of genetic and epigenetic factors, metabolism, and protein reactions at both the species and individual cell level. Development within a particular species and cell type has been shown to be determined by oscillations observed in the expression of keydevelopmental genes, with time scale of these processes also taking place at species-specific and cell-autonomous level. Individual genetic and epigenetic changes, circadian rhythmicity has been shown to persist at individual cell level in-vitro, with coupling of circadian clock and cell cycle resulting in timed mitosis, rhythmic DNA replication, and circadian and cell-cycle controlled transcription. The transcription-translation clock-controlled feedback loop drives expression of clock genes with a period of ~24hrs. These clock-controlled genes have been shown to comprise from 10-50% of a tissue’s transcriptome, varying in a tissue and cell-type dependent manner. Increasing evidence suggests the cell cycle, metabolic signaling, and circadian oscillators behave as a system of coupled oscillators. This synchronization challenges regulatory network inference, as correlated transcriptional changes can confound causal gene regulation with effect on cell cycle and circadian rhythm. This project’s goal is to determine how changes in dynamics of regulatory components, along regulatory networks are responsible for the regulation of the circadian, cell cycle, and metabolic oscillators; and to elucidate how these changes can lead to both cell fate decisions and disease development, e.g. cancer development. Determining components of the regulatory networks responsible for dysregulating synchronization of cells would be necessary. We will accomplish this by leveraging single cell-RNA Seq and spatial transcriptomic data available from a variety of developmental and cancer datasets to identify molecular components responsible for cell-fate determination with the goal of identifying how particular gene expression dynamics responsible for cell-fate determination are driven by cell cycle, circadian, metabolic oscillators. We will help identify these datasets, learn to process them to infer cyclic processes, and derive regulatory networks. We hypothesize that disease states will be characterized by a lack of synchronization and loss of direct regulation amongst cell-cycle, circadian, and metabolic oscillators and will be associated with a context-dependent regulatory network.

The COVID-19 pandemic underscored the importance of understanding the factors driving disproportionate health outcomes and integrating heterogeneity into epidemiological models. However, many existing models use homogeneous mixing assumptions, which fail to capture disparities within different populations. This project aims to address this gap by employing various quantitative risk assessment frameworks to analyze spatial heterogeneity in disease risk, considering sociodemographic and environmental factors. We will evaluate how different model parameters, disease characteristics, initial conditions, and data sources influence assessed risk. Our goal is to characterize risk variation both spatially and across social and demographic dimensions. The project will leverage techniques from epidemic modeling, spatial statistics, data science, and machine learning.

Controlling Bacterial Biofilm Growth


Bacterial biofilms are composed of bacterial cells along with matrix (self-produced polysaccharides, proteins, and DNA that are gel-like). Biofilms are dominant phenotype of bacteria and can be found in natural, industrial and medical settings. In the case of bacterial infections, we would want to prevent biofilm growth and spreading. In this project, we will utilize experimental data to develop agent-based models of bacterial biofilms. Conditions we could examine include the relationship between biofilm metabolism and mechanics, and/or the relationship between biofilm spreading and mechanics. We plan to assess different parameter estimation techniques, along with utilizing machine learning to identify potential rules that govern the agents in the system.

Agent-based modeling (ABM) is a computational method for analyzing and
predicting the emergent, population-level outcomes of interacting, autonomous
individuals in a complex system. ABM has been widely used to inform planning and
decision making across a variety of industries and sectors of society, including finance,
architecture and urban planning, national security and defense, sales and marketing,
social and political sciences, education, public health, medicine, and biomedical
research. In this project, participants will learn how to develop and code an ABM to
simulate the cellular and molecular mechanisms of human disease and to identify new
drug targets for treating disease. The goals of this project are to: 1) simulate human
fibrotic lung disease, 2) use ABM simulations to investigate the contributions of
fibroblast cellular heterogeneity to pathogenesis, and 3) use the ABM to identify novel
molecular targets for drugs that can slow or reverse disease progression. Participants
will learn how to simulate different initial conditions in order to explore the use of ABMs
as a framework for constructing a patient-specific “digital twin”. They will also use their
models to test real and hypothetical drugs for personalized medicine. Participants are
not required to have any prior experience with ABM. During this project, they will learn
how to program in a user-friendly, freely available ABM software called NetLogo.

Recent methodological advances have made it possible to infer tree-like genealogies that represent the genetic relationships within a sample of genomes from a population. Due to recombination, different regions of the genome may have distinct ancestral lineages, resulting in multiple genealogies. To better understand evolutionary processes, there is a need for methods that leverage these genealogies to infer key parameters. In this workshop, we will focus on developing computational methods that utilize genealogies to infer demographic events, such as introgression, and to detect positive selection. This will involve creating new approaches for comparing trees across the genome and developing statistical tools with the power to accurately infer the evolutionary processes of interest.