Fitness Landscapes and Surfaces of Selective Value

Abstract

The notion of “fitness landscapes” was presented by Sewall Wright in 1932. Its influence in evolutionary biology was extensive in several directions up to the present day. One direction consists in studies that built “fitness landscapes” although, according to my analysis, they employed only a part of Wright’s ideas - i.e. the one concerning “surfaces of selective value” (cf. Wright 1988) - focusing on one or few genetic or phenotypic traits of the studied systems. The model Wright fostered in 1932 was about the entire genotypic space of a Mendelian population, characterized by huge dimensionality. The lack of formal tools and computational power have prevented its actual construction, but understanding the original idea and how it differs from the realized models seems useful, all the most after the recent proposal by Sergey Gavrilets (e.g. Gavrilets 1997; 2004) of revising the overall structure of the genotype space. Understanding crucial differences is necessary here as well: for example, the newly proposed diagrams - namely, nearly flat, holey surfaces - do not represent the whole genotypic space, but the existence and properties of “nearly neutral networks” within it. The latter are fundamental for building particular speciation models called “spontaneous clusterization” (Gavrilets 2010). I will present, on the one hand, Wright’s primal proposal and the revision advanced by Gavrilets, on the other hand, the fruitful “surface of selective value” method, that consists in (1) representing genetic or phenotypic variants as points that are distributed on a bi-dimensional surface, so that the distance between points be proportional to the “reachability” between variants; (2) extruding such a surface along a third, orthogonal dimension that represents the considered variants’ fitness. The method aids the study of the role of fitness and other factors in evolutionary dynamics.