Deception and Diversity Dynamics in the Evolution of Drug Resistance in Malaria

Dr. Margaret Eppstein
Professor and Chair of Computer Science, University of Vermont

The fitness-landscape analogy has influenced the way scientists think about evolution for nearly a century. ‘Deception’ was coined in the evolutionary computation literature to describe landscapes in which sign epistasis can mislead a genetic algorithm away from the global optimum. However, current definitions of ruggedness and deception only capture intrinsic properties of landscapes. Here we quantify deception as a continuous function of the fitnesses and time-varying frequencies of genotypes, to reveal previously unexplored aspects of evolutionary dynamics for antimicrobial drug resistance. We use simulations to study deception dynamics in empirical and permuted landscapes for drug resistance in Plasmodium falciparum. Our results suggest a mechanism for how suboptimal variants of pathogens can dominate populations long after optimal mutants appear. We illustrate why certain drug regimens can promote the rapid evolution of variants that are both more drug resistant and more efficient replicators in the absence of drug. We also find that genetic diversity varies non-monotonically with different drugs, dosage levels, and duration of drug treatment – a finding with potential clinical relevance. Our approach can be readily applied to study other empirical fitness landscapes and toward an improved understanding of the evolution of antimicrobial drug resistance.
Pizza will be served @ 12:20 p.m.
This event is sponsored by the Computer Science Department