Regensburg 2025 – wissenschaftliches Programm

Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

BP: Fachverband Biologische Physik

BP 26: Synthetic life-like systems and Origins of Life

BP 26.9: Vortrag

Donnerstag, 20. März 2025, 12:00–12:15, H46

Linking fitness landscape topography to the characteristics of the underlying genotype-phenotype map — Malvika Srivastava^1,2, Ard A. Louis³, and •Nora S. Martin⁴ — ¹Institute of Integrative Biology, ETH Zurich, Zurich, Switzerland — ²Swiss Institute of Bioinformatics, Lausanne, Switzerland — ³Rudolf Peierls Centre for Theoretical Physics, University of Oxford, Oxford, UK — ⁴CRG (Barcelona Collaboratorium for Modelling and Predictive Biology), Dr. Aiguader 88, Barcelona 08003, Spain

The topographies of fitness landscapes are central in models of evolutionary processes. Key topographical features include the prevalence of fitness peaks, as well as the existence of accessible (i.e. fitness-increasing) paths to the global fitness optimum. Recent numerical work found that such accessible paths commonly exist in fitness landscapes based on biophysical models of genotype-phenotype (GP) maps, even when fitness values are randomly assigned to phenotypes [1]. Here, we examine such landscapes with random phenotype-fitness assignment more thoroughly to investigate, how their topography depends on the characteristics of the underlying GP map. By simplifying the GP map to a ``neutral component" (NC) graph, we can compute the expected prevalence of fitness peaks based only on two GP map characteristics: the evolvabilities and sizes of NCs. Evolvabilities are also important for peak heights and for the existence of accessible paths to global optima. [1] S. F. Greenbury, A. A. Louis, S. E. Ahnert, Nat Ecol Evol. 6, 1742-1752 (2022).

Keywords: fitness landscape; evolution; genotype-phenotype map; genotype network