EEB Tuesday Seminar Series - Toxic Relationships: The Genomic and Phenotypic Outcomes of Molecular Warfare between Predators and Prey// Applying Wafer-Scale Evolutionary Simulations to Investigate Hypermutator Dynamics in Large Asexual Populations

This event is part of our ongoing Tuesday Seminar Series.

-Talk title: Toxic Relationships: The Genomic and Phenotypic Outcomes of Molecular Warfare between Predators and Prey
Summary: A huge swathe of biodiversity interactions antagonistically via toxins. For example, some 15 percent of animal species use complex venoms for offense or defense. In turn, toxin resistance has evolved across the tree of life as well, making toxin-mediated interactions important subjects in the study of predator-prey coevolution. I will discuss ongoing work that unites comparative genome biology, population genomics, proteomics, and biochemical assays to determine the molecular targets of snake venoms and the broader evolutionary forces shaping venom variation among populations and species. I will next describe the observable phenotypic outcomes of coevolution between snakes and their venom resistant prey across landscapes. What emerges is a picture of how these traits may be shaped by a combination of biotic and abiotic forces to maintain biological diversity.

-Talk Title: Applying Wafer-Scale Evolutionary Simulations to Investigate Hypermutator Dynamics in Large Asexual Populations
Summary: Hypermutator alleles, which elevate mutation rates, introduce a fundamental trade-off in asexual organisms by accelerating adaptation but also increasing mutation load. In scenarios with rich opportunities for adaptive mutations, large population sizes can promote fixation of hypermutator strains via increased mutation supply. However, factors that might suppress hypermutator traits in large asexual populations are less understood. In this talk, I’ll discuss ongoing agent-based modeling experiments investigating how scenarios with few beneficial mutations available (i.e., limited adaptive potential) affect mutator dynamics across a continuum of population sizes and spatial structures. We will also survey technical aspects of the work, which harnesses the 850,000-processor Cerebras Wafer-Scale Engine to achieve simulation of population sizes ranging up to 1.5 billion with a 111,091× speedup over CPU execution. These experiments reflect an initial step in investigating the potential for next-generation computing hardware to substantially scale up scope tractable within in silico models. We will review further opportunities to develop foundational infrastructure that enables application of decentralized, many-processor simulations across diverse research agendas in evolutionary biology.

Join Remotely: https://umich.zoom.us/j/94285293918
Meeting ID: 942 8529 3918
Passcode: tuesem