Inferring fitness landscapes by regression produces biased estimates of epistasis

By Jakub Otwinowski, Joshua B. Plotkin. 2014

80 citations recorded by the source corpus.

Significance The dynamics of evolution depend on an organism’s fitness landscape: the mapping from genotypes to reproductive capacity. Knowledge of the fitness landscape can help resolve questions, such as how quickly a pathogen will acquire drug resistance or by what pattern of mutations. However, direct measurement of a fitness landscape is impossible because of the vast number of genotypes. Here, we critically examine regression techniques used to approximate fitness landscapes from data. We find that such regressions are subject to two inherent biases that distort the biological quantities of greatest interest, often making evolution seem less predictable than it actually is. We discuss methods that may mitigate these biases in some cases. The genotype–fitness map plays a fundamental role in shaping the dynamics of evolution. However, it is difficult to directly measure a fitness landscape in practice, because the number of possible genotypes is astronomical. One approach is to sample as many genotypes as possible, measure their fitnesses, and fit a statistical model of the landscape that includes additive and pairwise interactive effects between loci. Here, we elucidate the pitfalls of using such regressions by studying artificial but mathematically convenient fitness landscapes. We identify two sources of bias inherent in these regression procedures, each of which tends to underestimate high fitnesses and overestimate low fitnesses. We characterize these biases for random sampling of genotypes as well as samples drawn from a population under selection in the Wright–Fisher model of evolutionary dynamics. We show that common measures of epistasis, such as the number of monotonically increasing paths between ancestral and derived genotypes, the prevalence of sign epistasis, and the number of local fitness maxima, are distorted in the inferred landscape. As a result, the inferred landscape will provide systematically biased predictions for the dynamics of adaptation. We identify the same biases in a computational RNA-folding landscape as well as regulatory sequence binding data treated with the same fitting procedure. Finally, we present a method to ameliorate these biases in some cases.

Subfields: Genetics, Sociology and Political Science

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Genetics

In Sociology and Political Science

Citations in the corpus

Coevolutionary Landscape Inference and the Context-Dependence of Mutations in Beta-Lactamase TEM-1 (2015, background): “been argued that existing mutagenesis data are not sufficient for accurate landscape regression [24].”
Adaptation in protein fitness landscapes is facilitated by indirect paths (2016, methodology): “We used predictor selection based on biological knowledge, followed by regularized regression, which has been demonstrated to ameliorate possible bias in the inferred fitness landscape (Otwinowski and Plotkin, 2014) .”
How Pairwise Coevolutionary Models Capture the Collective Residue Variability in Proteins? (2018, background): “[33] for the related case of inferring epistatic fitness landscapes.”
Inferring the shape of global epistasis (2018, methodology): “The global epistasis model also makes reasonable predictions for genotypes far from those sampled in the mutagenesis assay, unlike the qualitatively incorrect out-of-sample predictions exhibited by pairwise models (10, 11).”
Networks beyond pairwise interactions: structure and dynamics (2020, background): “[701] Jakub Otwinowski and Joshua B Plotkin.”