What is Adaptive Landscape and how to study it?

What is adaptive Landscape?

An adaptive landscape is a surface in multidimensional space (analogous to a mountain range) that represents the mean fitness of a population (not the fitness of a genotype). An individual is represented as a point on the surface (mountain) and a population as a cloud.

Adaptive landscapes occupy a central and special position in the theory of evolution, especially in population and quantitative genetics, but also in some models of macroevolution. The adaptive landscape concept was originally formulated by the population geneticist Sewall Wright in his now-classic work from 1932 entitled “The role of mutation, inbreeding, and crossing in evolution”.

An adaptive landscape shows the relationship between fitness (vertical axis) and one or more traits or genes (horizontal axis). An adaptive landscape is therefore a form of the response surface that describes how a dependent variable (fitness) is causally influenced by one or more predictor variables (trait or genes).

Evolution through natural selection in the context of adaptive landscapes can be viewed as a mountaineering process in which populations ascend to the trait or gene combination with the highest fitness (“adaptive peaks”).

Between the adaptive peaks, there are regions in the phenotype or genotype space with low fitness (“fitness valleys”). Adaptive landscapes have appeared in popular science literature on evolutionary biology, perhaps most explicitly in Richard Dawkins’ 1996 book Climbing Mount Improbable.

Adaptive landscape

Here, like many others, Dawkins compared evolution through natural selection to a mountaineering process in which populations and species evolved by climbing the closest adaptive peak, although this is not always the peak with the highest overall fitness.

In the pre-computer era, adaptive landscapes were often represented as three-dimensional surface diagrams with only two features (or genes) as independent variables. Adaptive landscapes then mainly had a heuristic function for qualitative thinking and not for quantitative analysis. With the fast

With the development of evolutionary quantitative gene theory and increasing computer power, adaptive landscapes are now one of several analytical tools for modern evolutionary biologists. However, the adaptive landscape concept has also generated a lot of criticism from both philosophers and evolutionary biologists

Wright’s adaptive landscapes.

Early in his career, Wright concluded from his work with the animal breeding program that the interaction between loci (epistasis) is common and that individual character can be influenced by a number of genetic factors (pleiotrophy).

He views evolution as a selection process in a network of interdependent genetic factors rather than in individual loci with independent effects, a view supported by R.A. Fisherman. With thousands of loci, the assumption of strong genetic interactions naturally leads to the conclusion that there must be multiple fitness optima, each of which represents a unique genetic combination.

Hence, the epistasis creates a rugged adaptive landscape with multiple peaks and valleys as opposed to a single fitness optimum that would be expected if all combinations of loci were purely additive.

While a two-dimensional projection is insufficient to represent such a complex multidimensional genotypic space, Wright’s view of an adaptive landscape has served as an important heuristic tool for understanding evolutionary processes.

Analytical Approaches and Methods of Study to Adaptive Landscape

There are five main traditional empirical approaches to studying adaptive landscapes:

  1. Measurement and quantification of the type of selection in natural populations.
  2. Experimental manipulations of individual genotypes, phenotypic traits or means of selection.
  3. Experimental evolution.
  4. Phylogenetic comparative methods.
  5. Conclusions from time series in the fossil record.

All of these methods have their advantages and disadvantages. Methods 1, 4, and 5 are correlative approaches and inferential, while Methods 2 and 3 manipulate either phenotypes, selective environments, or both.

The analysis and visualization of adaptive landscapes and fitness surfaces has become more sophisticated with the development of multivariate statistical techniques.

Today, evolutionary biologists can describe fitness surfaces in terms of both height and curvature and have developed a rigorous formal mathematical framework to quantify adaptive peaks and fitness troughs, as well as interactions between features (“correlational selection”) and their effects on fitness.

The development of multivariate statistical methods for analyzing fitness surfaces now means that evolutionary biologists are no longer forced to think about adaptive landscapes in only three dimensions, which has been a common criticism of adaptive landscapes.

The parameters used to quantify fitness surfaces include directional selection gradients (β), curvilinear, and correlative selection gradients. Fitness is then modeled as a response area, which is represented using one or more underlying phenotypic characteristics (predictor variables).

Different Types of Adaptive Landscapes

Landscape Structure

Various types of adaptive landscapes have been developed since Wright, the development of which will follow very different dynamics. Wrightian adaptive landscapes are characterized by multiple adaptive peaks and intermediate fitness troughs that reflect the fitness epistasis expected from strong gene interactions.

“Fisherian Adaptive Landscapes” are smoother than their Wrightian counterparts and are characterized by a single adaptive peak that the population moves to through the action of mass selection in large and panmic populations.

In contrast, in the “harsh adaptive landscapes” envisaged by Wright, populations remain stuck on local adaptive peaks for a very long time and cannot move across fitness valleys. In “holey adaptive landscapes”, evolution can easily run along neutral or semi-neutral ridges that connect genotypes of roughly equal fitness throughout the genotype space.

The target model of adaptation is an application of adaptive landscape theory originally proposed by Fisher and formally modeled by Orr to understand the genetic architecture of the adaptation process through natural selection towards a fixed optimum.

The target model predicts an exponential size distribution of the fixed factors involved in the adaptation, with genes with large effects typically being pinned first, followed by genes with smaller effects as the population approaches optimum.

The reason for this sequence of events that lead to the fixation of factors of different sizes is that genes with great effect at the beginning of the population history tend to bring the population closer to the fitness optimum, but as the optimum approaches, mutations with great effect result rather to an “overshoot” of the optimum and thus reduce fitness.

Landscape Dynamics

Natural and sexual selection in natural populations often varies widely when measured over different years or generations. These fluctuating selection pressures strongly suggest that adaptive peaks and valleys are not stable on microevolutionary time scales.

In addition, adaptive peaks and fitness valleys are expected to be even less stable on macroevolutionary timescales due to large climatic and biotic changes. The ecological causes and evolutionary consequences of such landscape dynamics are largely unknown and are widely discussed by evolutionary biologists.

Thus, a distinction can be made between “static” and “dynamic” adaptive landscapes. Substantial changes in the environment, such as moving into new “adaptive zones”, are generally considered to be the main explanation for peak movements on macroevolutionary timescales, as envisaged by Simpson.

On microevolutionary time scales, density- and frequency-dependent selection as mechanisms that cause landscape dynamics and the movement of adaptive peaks has attracted particular interest. For example, a rare morph or genotype with negative frequency-dependent selection (NFDS) experiences high fitness due to its low frequency, but this fitness advantage disappears when it increases in frequency and climbs the next adaptive peak.

Therefore, a fitness peak becomes a fitness trough over time, and a fitness trough becomes a fitness peak over generations as the genetic makeup of the population changes. The adaptive landscape is therefore more like an “ocean” with moving waves than a static mountain range.

Other ecological interactions and processes such as coevolutionary arms races and the “Red Queen” dynamics can have similar effects, and Fisher emphasized this aspect of the evolutionary process in his “Fundamental Theorem of Natural Selection.”

Therefore, under NFDS and similar ecological scenarios, although the fitness of some genotypes does increase, the mean fitness of the population will not necessarily increase, indicating the crucial distinction between the fitness surface of individuals and the mean fitness of the population.