MCMC Seminar Speakers

Disentangling correlation and causation in somatic evolution

Kamila Naxerova

(Upcoming) April 7, 2026, 11 a.m. - 12 p.m. PT

Somatic evolution—the interplay of mutation, selection, drift, and cell division—shapes cell populations across human tissues. Increasingly, it is being quantified using evolutionary and statistical frameworks, yet a central challenge remains: distinguishing whether observed evolutionary patterns are drivers of disease or consequences of disease exposure. This talk presents two examples that highlight this problem. First, disease-driven changes in tissue dynamics, such as altered stem cell proliferation, can modulate the rate of somatic evolution and thereby accelerate the expansion of mutant clones. This provides a mechanistic explanation for associations between clonal expansions and disease that must be considered as an alternative to the notion that mutant clones cause pathology. Second, a quantitative framework for separating selection from causation in cancer genomes is introduced. Cancer genomes are enriched for mutations that cause cancer, but they also integrate signals from decades of normal tissue evolution that are passively inherited by malignant lineages. By integrating evolutionary modeling with patient-level data, including age distributions, this approach enables estimation of mutations’ contributions to cancer risk independently of their selective advantage in normal tissue. Together, these results highlight a general principle: somatic evolution is both shaped by and contributes to disease, and quantitative approaches can help with the difficult problem of disentangling correlation from causation.

Mathematical modeling of cancer evolution to optimize prevention strategies

Kit Curtius

September 9, 2025

Seminar Recording Link

Cancer initiation and progression is an evolutionary process. Genetic and epigenetic alterations underpin phenotypes that drive natural selection at the cellular level in precancerous stages. However, predicting later cancer formation often requires more than counting mutations. Understanding the dynamics is critical to predict individual cancer risk in patients and intervene effectively. I will present mathematical methods that incorporate multiscale data types (from population-level incidence to patient-level genotypes) to help answer complex questions such as when to offer screening tests in certain populations.

Modeling tumor heterogeneity to understand the clinical efficacy of combination therapy

Adam Palmer

May 13, 2025

Seminar Recording Link

Tumor heterogeneity is a central difficulty of cancer treatment and causes practically all cancer therapies to vary in effectiveness between patients. Combination therapy is a key strategy to overcome heterogeneity, but predicting which drug combinations will be effective has been a long unsolved problem. Because single drugs are variably effective, we have built models around the idea that combination therapy in human populations involves ‘variable effects plus variable effects’. I will describe studies that apply this principle to understand and to accurately predict the population-level activity of combination therapies in human clinical trials, including treatments with curative effect.

Understanding, predicting and controlling stochastic cancer evolution

Rob Noble

March 4, 2025

Seminar Recording Link

Understanding the nature of tumour evolution promises to enable more accurate prognostic methods and more effective treatment strategies. I will use three examples to illustrate how the analysis of stochastic processes can aid this goal by bridging the gap between ODE/PDE models and agent-based simulations. First, I will show how surprisingly simple mathematical expressions can be derived to explain why selective sweeps (the spread of beneficial mutations through an entire population) are rare except when tumours are relatively very small. Next, I will explain how studying tree generating processes and tree shape indices can improve model selection and clinical forecasting methods. Finally, I will present an application of stochastic processes to improving cancer cure rates by minimizing the probability of evolutionary rescue. Although all this work is motivated by questions in cancer research, the methods and results are readily applicable to other biological systems.