2015/07/29 by François Blanquart, Blanquart, François, Thomas Bataillon +1 · 1 citation
Biochemistry, Genetics and Molecular Biology · Social Sciences · #Evolution and Genetic Dynamics #Evolutionary Game Theory and Cooperation #FOS: Biological sciences #Genetic diversity and population structure #Populations and Evolution (q-bio.PE)
paper · pdf · doi:10.48550/arxiv.1507.08245
openalex publication_date 2015/07/29 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
The fitness landscape defines the relationship between genotypes and fitness\nin a given environment, and underlies fundamental quantities such as the\ndistribution of selection coefficient, or the magnitude and type of epistasis.\nA better understanding of variation of landscape structure across species and\nenvironments is thus necessary to understand and predict how populations will\nadapt. An increasing number of experiments investigates the properties of\nfitness landscapes by identifying mutations, constructing genotypes with\ncombinations of these mutations, and measuring the fitness of these genotypes.\nYet these empirical landscapes represent a very small sample of the vast space\nof all possible genotypes, and this sample is often biased by the protocol used\nto identify mutations. Here we develop a rigorous statistical framework based\non Approximate Bayesian Computation to address these concerns, and use this\nflexible framework to fit a broad class of phenotypic fitness models (including\nFisher's model) to 26 empirical landscapes representing 9 diverse biological\nsystems. In spite of uncertainty due to the small size of most published\nempirical landscapes, the inferred landscapes have similar structure in similar\nbiological systems. Surprisingly, goodness of fit tests reveal that this class\nof phenotypic models, which has been successful so far in interpreting\nexperimental data, is a plausible model in only 3 out of 9 biological systems.\nMore precisely, although Fisher's model was able to explain several statistical\nproperties of the landscapes - including mean and standard deviation of\nselection and epistasis coefficients -, it was often unable to explain the full\nstructure of fitness landscapes.\n