Table 1.

Summary of important conceptual advances in the study of G→P mapping. References are meant to be useful for further readings, and they are by far not exhaustive.

research area	conceptual advance	references
computational science	modularity, robustness and evolvability are tightly related, with an intermediate degree of modularity allowing for a compromise of high robustness and high evolvability (the two are otherwise negatively related)	Ciliberti et al. (2007), Gjuvsland et al. (2007) and Crombach & Hogeweg (2008)
computational science	G→P functions evolve to increase evolvability, which is made possible by the ‘hub-and-spoke’ structure of genetic networks	Crombach & Hogeweg (2008)
computational science	robustness arises because of the general properties of feedback systems that are built into gene networks; this means that natural selection can indirectly affect robustness	Gjuvsland et al. (2007)
computational science	the large neutral or quasi-neutral areas of genotypic space that are necessary for the evolution of robustness/evolvability are intrinsic properties of highly dimensional genotypic spaces	Gravner et al. (2007)
computational science	artificial development, used to generate ‘evolvable hardware,’ provides a good model and theoretical foundation for why developmental encoding evolved on top of genetic encoding	Hartmann et al. (2007) and Roggen et al. (2007)
gene networks	a gene's position in a network determines its evolutionary role; genes are not equivalent entities as in standard population genetic models	Stumpf et al. (2007), Chouard (2008) and Stern & Orgogozo (2009)
gene networks	evolution can turn a gene network into another by altering the regulation of individual components while leaving the ‘logical output’ unchanged	Tsong et al. (2006)
general theory	introduction of G→P mapping concept	Alberch (1991)
general theory	properties of mutational one-step networks determine evolutionary paths	Kauffman & Levin (1987)
general theory	G→P mapping results in punctuated equilibria patterns	Cowperthwaite & Meyers (2007)
general theory	large neutral areas of genotypic landscape allow for evolutionary search of novel phenotypes (evolvability)	Cowperthwaite & Meyers (2007), Fernández & Solé (2007), Sumedha et al. (2007) and Wroe et al. (2007)
general theory	developmental encoding allows for the evolution of complex phenotypes by decoupling the proportionality between length of genetic instructions and complexity of phenotypes; the evolutionary search space becomes smaller; in this sense, development is the G→P map	Ciliberti et al. (2007) and Stanley (2007)
RNA folding	RNA folding is an empirically tractable model of G→P mapping	Fontana (2002)
RNA folding	G→P mapping is characterized by high degrees of ‘degeneracy’ or genetic redundancy, where the same phenotype is generated by a variety of not necessarily similar genotypes	Stich et al. (2008)
RNA folding	connectivity in genotypic space is asymmetrical, so evolution depends to some extent on contingent events	Cowperthwaite & Meyers (2007)
RNA folding	genotypic one-step neighbours are often connected to phenotypes that are not structurally similar	Sumedha et al. (2007)
RNA folding	‘species’ and specialized ecological functions (including parasitism) evolve spontaneously as emergent properties of populations of replicators	Takeuchi & Hogeweg (2008)