Epistasis and the structure of fitness landscapes: are experimental fitness landscapes compatible with Fisher's Geometric model?

TL;DR

This study develops a Bayesian framework to analyze empirical fitness landscapes and tests the compatibility of Fisher's Geometric model with data from diverse biological systems, revealing limited applicability.

Contribution

Introduces a statistical framework to rigorously fit phenotypic fitness models to empirical landscapes, challenging the universality of Fisher's model across systems.

Findings

Fisher's model fits only 3 out of 9 systems.

Empirical landscapes show similar structure within systems.

Fisher's model explains some statistical properties but not full landscape structure.

Abstract

The fitness landscape defines the relationship between genotypes and fitness in a given environment, and underlies fundamental quantities such as the distribution of selection coefficient, or the magnitude and type of epistasis. A better understanding of variation of landscape structure across species and environments is thus necessary to understand and predict how populations will adapt. An increasing number of experiments investigates the properties of fitness landscapes by identifying mutations, constructing genotypes with combinations of these mutations, and measuring the fitness of these genotypes. Yet these empirical landscapes represent a very small sample of the vast space of all possible genotypes, and this sample is often biased by the protocol used to identify mutations. Here we develop a rigorous statistical framework based on Approximate Bayesian Computation to address these concerns, and use this flexible framework to fit a broad class of phenotypic fitness models (including Fisher's model) to 26 empirical landscapes representing 9 diverse biological systems. In spite of uncertainty due to the small size of most published empirical landscapes, the inferred landscapes have similar structure in similar biological systems. Surprisingly, goodness of fit tests reveal that this class of phenotypic models, which has been successful so far in interpreting experimental data, is a plausible model in only 3 out of 9 biological systems. More precisely, although Fisher's model was able to explain several statistical properties of the landscapes - including mean and standard deviation of selection and epistasis coefficients -, it was often unable to explain the full structure of fitness landscapes.