Midwest Mathematical Biology Seminar

Schedule (Spring 2026)

(All talks at 11am Eastern Time / 10am Central Time on Tuesdays)

Zoom link:  https://illinois.zoom.us/j/84907159401?pwd=JboBtq30vntVuEpj8s3yYfAdwlb258.1

• January 13: Maximillian Newman (University of Chicago)

Title: How to model genetic inheritance across the genome

Abstract: Statistics in population genetics rely on understanding how a sample n genes in a population of size N >> n coalesce backwards in time to form a genetic tree called a gene genealogy, a tree-like structure that encodes the relatedness of the samples. Implicit in population genetic methods is the assumption that genes sampled far enough apart on the genome have independent genealogies. I will explain how in structured populations with large migrations and uneven offspring distributions this independence assumption fails, and what new mathematical object arise when one tries to model the distribution of these genealogies across the genome. This is based on probabilistic work, though I will assume minimal background in probability theory.

• January 20: Open

• January 27: Haridas Das (Oklahoma State University)

Title: Exploring Disease Transmission in Metapopulations: The Influence of Human Mobility and Heterogeneity

Abstract: Human mobility is a key driver of disease transmission, shaping how outbreaks unfold across regions. In this talk, I will show how mobility-driven networks generate distinct disease dynamics in metapopulations, which are geographically separated subpopulations connected through movement. In the first part, we use a metapopulation framework to analyze how network structure affects the basic reproduction number (R₀). We introduce the Standard Threshold Property and identify network classes, such as fully connected and star-shaped networks, that share epidemic thresholds, while cycle-shaped networks exhibit distinct dynamics. The second part integrates cellphone-based mobility data into a county-level metapopulation model across the United States. We identify hotspot and superspreader counties and evaluate targeted interventions, including localized R₀ reduction and selective mobility restrictions. Results show that hybrid strategies focusing on key counties can achieve epidemic control comparable to statewide lockdowns. These studies highlight how combining network theory with mobility data can inform adaptive, effective strategies for disease control and pandemic preparedness.

• February 3: Open

• February 10: Open

• February 17: Open

• February 24: John Metzcar (University of Minnesota)

Title: GBM-FORECAST: Predicting recurrence in glioblastoma using serial white blood cell counts and patient advocacy influences on my career

Abstract: Glioblastoma is an aggressive cancer with poor progression-free survival. Rapid recurrence detection is critical to enhance the benefits of secondary treatment. White blood cell (WBC) counts and WBC-derived markers are known recurrence biomarkers, but prior work has treated them as static, single risk assessments, often using data far from recurrence. However, WBC counts are not static and are measured throughout follow-up and multiple studies show changes in immune cell population dynamics driven both by treatments and tumor progression. In this work, we consider the potential of longitudinal WBC data to predict recurrence. We also consider their potential as dynamic recurrence risk biomarkers. Training random survival forest models to predict recurrence, we found low prediction error as measured by the integrated Brier score and moderate ability to distinguish early and late recurring patients as measured by the integrated dynamic AUC, supporting the presence of informative signal in the data. Moving to model application, we calculated a near-term, recurrence risk score, the GBM-FORECAST score. Despite the absence of imposed temporal structure, we observed quantitative cross-patient risk score patterns: an initial increase, followed by a decline, and subsequent increase in score through the end of follow-up. These transition points coincide with the shift from chemoradiotherapy to maintenance chemotherapy and the onset of increasing recurrence rate. While the origin of these patterns requires further study, they suggest a connection between counts, treatments, and disease dynamics. Furthermore, this analysis may be able to quantitatively connect unobservable tumor microenvironment processes to easily obtained clinical measurements, opening up opportunities for therapy design through virtual clinical trials. I will also talk about aspects of my career related to a cancer diagnosis and patient advocacy.

• March 3: Longhua Zhao (Case western reserve)

Title: Fluid-structure interaction: modeling, applications and a practical guide for undergraduate researchers

Abstract:

Mathematical and computational modeling plays an important role in understanding the mechanics of life in low-Reynolds-number flows. The method of regularized Stokeslets (MRS) is a popular approach for solving fluid–structure interaction problems in this regime. In this talk, I present the development of an introductory guide to MRS designed to mentor undergraduate research students, along with research results utilizing the method in applications such as cilia, helical flagella, and microfluidic tweezers.

• March 10: Rui Wang (New York University)

Title: Applied Topology, Topological Spectral Theory, and Machine Learning Models for Biological Applications

Abstract: Topological methods provide powerful tools for analyzing complex, high-dimensional data. In this talk, I will present recent advances in Topological Deep Learning (TDL) and their applications to key problems in bioscience. A central concept in TDL is Persistent Spectral Graphs (PSGs), a method in Topological Data Analysis (TDA) that is designed to capture intricate topological changes and homotopy shape evolution information for high-dimensional biological data. I will demonstrate how PSG-based representations of biomolecules can be incorporated into AI pipelines to yield TDL models with strong predictive power. Applications span multiple biological domains, including forecasting the emergence of new COVID-19 variants, inferring RNA secondary structure motifs, and improving model selection for peptide–protein complexes predicted by AlphaFold2 and AlphaFold3.

• March 17: Artem Novozhilov (North Dakota State University)

• March 24: Abdel Halloway (Case Western Reserve University)

• March 31: Harman Jaggi (Princeton University)

• April 7: Desmond Yengi (Grinnell College)

• April 14: Nour Khoudari (Purdue University)

• April 21: Andrew Krause (Durham University)

• April 28: Changhan He (University of South Carolina)

• May 5: Naoki Masuda (University of Michigan)

• May 12: Kyle Dahlin (Virginia Tech)

• May 19: Uduak George (San Diego State University)

• May 26: Wanda Strychalski (Case Western Reserve University)