Página Inicial » Inscrições Científicas » Trabalhos

Dados do Trabalho




Most of our knowledge of the relationship between modularity and the adaptive landscape comes from simulation studies. In simulations carried out on simple landscapes, patterns and magnitudes of association among traits (G-matrix) affect the direction, magnitude, and rate of evolutionary change under selection. The effect of the G-matrix on evolutionary change depends critically on its structure in relation to the adaptive landscape and can either augment or slow the evolutionary response relative to a situation with fully independent traits. If selection is along dimensions unaligned with modularity/integration patterns, the response is deflected toward the lines of least resistance. If selection is aligned with modularity, however, the evolutionary response is greatly facilitated. The closer the alignment with the major line of least resistance, the quicker and more direct the evolutionary response. However, simulations are highly concordant in showing that these effects are restricted to the microevolutionary scale, and, given sufficient time and a simple adaptive landscape, the population will eventually reach the selective peak, unless there is no genetic variation at all in that direction (an absolute constraint). But theoretical work suggests that even if there is an apparent lack of genetic variation along some dimension, genetic variation is frequently hidden in the form of epistasis that can fuel evolutionary change in subsequent generations. Therefore, given the possibility of adaptive changes in the G-matrix through time and the understanding that constraints imposed by G-matrices are usually microevolutionary, the emerging picture would be one in which G-matrices should not have any enduring macroevolutionary consequences. But what happens when we consider complex adaptive landscapes? Although single-peaked adaptive landscapes are convenient for model building purposes, adaptive landscapes are thought to be very rugged, that is, they have many adaptive peaks and valleys. When the adaptive landscape is rugged and genetic associations are stable through time, macroevolutionary dynamics are shaped by the interaction between the G-matrix and the adaptive landscape. This interaction implies that, in rugged and multiple-peaked adaptive landscapes, the G-matrix can have a major influence in determining which peak will be reached by a given population, even if in theory the effect of the G-matrix is microevolutionary. This argument can be easily understood noting that, in evolutionary terms, the distance between the population average position and the peak is not a simple linear (Euclidean) distance between the start position and end position of the species averages but is a weighted distance, with the weight being given by the patterns of genetic association. Given the influence of genetic correlations, the distance of a population from a peak is measured in units of genetic variation. Thus, the closest peak, the peak the population eventually reaches, is not necessarily the highest or even the closest in Euclidian distance but is the closest in genetic-scaled distance. We contrast both models (simple x rugged landscape) in terms of expected macroevolutionary patterns while presenting empirical findings in mammals. Mammals present a strong relationship between the alignment of evolutionary change with G-matrix and the magnitude of evolution.


Adaptive Landscape, Constraints, Micro and macroevolution, 






Gabriel Marroig