Mutational robustness (persistence of a phenotype under mutation) and evolvability (capacity to produce heritable phenotypic variation) are antagonistic at the genotype level, positively associated at the phenotype level in RNA secondary-structure maps, measure-dependent in gene-network evolution models (evolvability has been reported alongside increased viability robustness but decreased phenotypic robustness), and simultaneously antagonistic and synergistic at different organizational levels in transcription-factor binding-site data. Wagner (2008) explicitly left open how general the positive phenotype-level association is. Subsequent work delineated the conditions for sign flips only within disjoint model classes — population-genetic models of adaptation speed, random discrete genotype–phenotype maps, Boolean hierarchies, digital organisms — under mutually incompatible assumptions and measures, and the most recent dedicated reviews (2025) still describe the impact of robustness on evolvability as debated. It remains open how generally the positive phenotype-level association demonstrated for RNA holds in other biological genotype–phenotype maps: an internally consistent account that, applied to a system with stated assumptions, states whether mutational robustness there promotes or retards evolvability, and through which mechanism, as an explicit function of level of description, map class, the statistical population compared, and evolutionary regime.

Background

Every living system can be described by a genotype–phenotype (GP) map: the rule that assigns phenotypes — molecular structures, gene-expression patterns, quantitative traits — to genotypes such as nucleic-acid sequences or regulatory architectures. Two properties of GP maps sit in a famous tension. Mutational robustness is the degree to which a phenotype persists unchanged when its genotype is mutated. Evolvability is a system's capacity to produce heritable phenotypic variation on which selection can act. At face value they are opposed: a system that buffers mutations produces less visible variation per mutation. Yet genotypes producing the same phenotype are often connected into neutral networks — sets of genotypes reachable from one another by single mutations that each preserve the phenotype — and such networks simultaneously imply robustness (many mutations are neutral) and a route to novelty (many different phenotypes may lie one mutation away from some member of the network). Cryptic genetic variation — genetic variation whose phenotypic effects are buffered until the environment or genetic background changes — is the experimental face of the same buffering. Wagner (2008) resolved the tension for one computational map: RNA molecules of length 30 nucleotides folded by minimum-free-energy rules into secondary structures (their intramolecular base-pairing patterns). The resolution distinguishes two levels of description. Genotype-level robustness (the fraction of a sequence's one-mutant neighbors that fold to the same structure) correlates negatively with genotype-level evolvability (the number of distinct structures in that one-mutation neighborhood; Spearman rank-correlation coefficient s ≈ −0.64), as had earlier been observed for RNA by Ancel and Fontana (2000). Phenotype-level robustness (neutrality averaged over all genotypes with a given structure) correlates positively with phenotype-level evolvability (the number of distinct structures reachable within one mutation of anywhere on that phenotype's neutral network; Spearman s ≈ 0.55). The mechanism is that phenotype frequencies — and hence neutral-network sizes — vary over orders of magnitude while genotype-level neutrality varies only modestly, and the one-mutation neighborhoods of different sequences on the same network contain mostly different structures; populations evolving on large networks therefore accumulate more genetic diversity and encounter more phenotypic variety, in both large and small population-mutation regimes. Wagner explicitly left open whether the positive phenotype-level association holds in other biological systems. Later work sharpened rather than settled this generality question. In a gene-network model — developmental regulatory circuits whose phenotype is a stable gene-expression pattern — Draghi and Wagner (2009) showed that fluctuating selection can increase evolvability and revealed mechanisms connecting evolvability, genetic architecture, and robustness. In a related gene-network model under autocorrelated fluctuating selection, Steiner (2012) found the sign to be measure-dependent: evolvability was correlated with increasing genetic robustness to effects on network viability and decreasing robustness to effects on phenotypic expression. For transcription-factor (TF) binding sites in mouse and yeast, Payne and Wagner (2014) found antagonism between robustness and evolvability at the level of individual binding sites coexisting with synergy at the level of binding-site families organized into large connected genotype sets; high-affinity sites tolerate the most substitutions. In digital organisms (self-replicating computer programs evolving in the Avida platform), Elena and Sanjuán (2008) found the effect of robustness on evolvability to be time-scale dependent in simple environments — neutral or counterproductive over short horizons, favorable over long ones — a dependence they present as compatible with the earlier RNA results. Reviews (Masel and Trotter 2010; Payne and Wagner 2019) consolidated this landscape without resolving when the sign flips.

Problem Statement

An internally consistent account is sought that determines how generally the positive association between phenotype-level mutational robustness and phenotype-level evolvability demonstrated for RNA secondary-structure maps by Wagner (2008) holds in other biological genotype–phenotype maps, and what fixes its sign. Wagner left this open verbatim: "how general is the positive association between phenotypic robustness and evolvability that I showed for RNA structures? Does it occur in many other biological systems?" The account should state, as an explicit function of: (a) the level of description — genotype-level neutrality and genotype-level neighborhood evolvability versus phenotype-level (neutral-network averaged) robustness and phenotype-level evolvability, in Wagner's sense; (b) the class of genotype–phenotype map (e.g., RNA secondary structure, protein-structure/lattice-folding models, transcription-factor binding sites, gene-regulatory-network models — which Wagner named as an important task for future work — digital organisms, random discrete maps); (c) the statistical population actually being compared — among phenotypes within one map versus across maps — and the scale on which phenotypes are measured (discrete, ordinal, continuous), since these can yield opposite signs; and (d) the evolutionary regime (population size and mutation rate, i.e. large versus small population-mutation regimes, time horizon, mode and rate of environmental change, recombination) — whether mutational robustness in a given system promotes or retards evolvability, and through which mechanism. Applied to a new system with stated assumptions, the account should tell a qualified domain expert the expected sign and mechanism there. Because robustness and evolvability have several non-equivalent operational definitions across these literatures, the account must either fix a common formalization or establish explicit translations among the measures it compares. Since the question asks whether the association occurs in other biological systems, the account's claims about a given system should be confronted with the evidence available for that system. Known results with which the account must be consistent, or which it must explicitly reconcile, include: the two-level RNA result (genotype-level antagonism, phenotype-level synergy); positive within-map but negative across-map correlations in random Boolean genotype–phenotype maps, with the positive relation failing for ordinal or higher-scale phenotypes (Mayer & Hansen 2017); the short- versus long-horizon dependence in digital organisms in simple environments, presented by its authors as compatible with the RNA results (Elena & Sanjuán 2008); the synergy at the level of transcription-factor binding-site families organized into large connected genotype networks (Payne & Wagner 2014); and the measure-dependent sign in evolving gene networks under fluctuating selection (Steiner 2012). An answer that re-derives the relationship within a single model class, or that redescribes the known contrasts without identifying the conditions under which the sign differs, does not answer this generality question.

References

  1. Wagner, A. (2008). Robustness and evolvability: a paradox resolved. Proceedings of the Royal Society B 275:91–100. DOI: 10.1098/rspb.2007.1137. https://doi.org/10.1098/rspb.2007.1137
  2. Draghi, J. A. & Wagner, G. P. (2009). The evolutionary dynamics of evolvability in a gene network model. Journal of Evolutionary Biology 22:599–611. DOI: 10.1111/j.1420-9101.2008.01663.x. https://doi.org/10.1111/j.1420-9101.2008.01663.x
  3. Payne, J. L. & Wagner, A. (2014). The robustness and evolvability of transcription factor binding sites. Science 343:875–877. DOI: 10.1126/science.1249046. https://doi.org/10.1126/science.1249046
  4. Ancel, L. W. & Fontana, W. (2000). Plasticity, evolvability, and modularity in RNA. Journal of Experimental Zoology 288:242–283. DOI: 10.1002/1097-010X(20001015)288:3<242::AID-JEZ5>3.0.CO;2-O. https://doi.org/10.1002/1097-010X(20001015)288:3%3C242::AID-JEZ5%3E3.0.CO;2-O
  5. Elena, S. F. & Sanjuán, R. (2008). The effect of genetic robustness on evolvability in digital organisms. BMC Evolutionary Biology 8:284. DOI: 10.1186/1471-2148-8-284. https://doi.org/10.1186/1471-2148-8-284
  6. Draghi, J. A., Parsons, T. L., Wagner, G. P. & Plotkin, J. B. (2010). Mutational robustness can facilitate adaptation. Nature 463:353–355. DOI: 10.1038/nature08694. https://doi.org/10.1038/nature08694
  7. Masel, J. & Trotter, M. V. (2010). Robustness and evolvability. Trends in Genetics 26:406–414. DOI: 10.1016/j.tig.2010.06.002. https://doi.org/10.1016/j.tig.2010.06.002
  8. Steiner, C. F. (2012). Environmental noise, genetic diversity and the evolution of evolvability and robustness in model gene networks. PLOS ONE 7:e52204. DOI: 10.1371/journal.pone.0052204. https://doi.org/10.1371/journal.pone.0052204
  9. Greenbury, S. F., Schaper, S., Ahnert, S. E. & Louis, A. A. (2016). Genetic correlations greatly increase mutational robustness and can both reduce and enhance evolvability. PLOS Computational Biology 12:e1004773. DOI: 10.1371/journal.pcbi.1004773. https://doi.org/10.1371/journal.pcbi.1004773
  10. Mayer, C. & Hansen, T. F. (2017). Evolvability and robustness: a paradox restored. Journal of Theoretical Biology 430:78–85. DOI: 10.1016/j.jtbi.2017.07.004. https://doi.org/10.1016/j.jtbi.2017.07.004
  11. Wei, X. & Zhang, J. (2017). Why phenotype robustness promotes phenotype evolvability. Genome Biology and Evolution 9:3509–3515. DOI: 10.1093/gbe/evx264. https://doi.org/10.1093/gbe/evx264
  12. Payne, J. L. & Wagner, A. (2019). The causes of evolvability and their evolution. Nature Reviews Genetics 20:24–38. DOI: 10.1038/s41576-018-0069-z. https://doi.org/10.1038/s41576-018-0069-z
  13. Jiang, P., Kreitman, M. & Reinitz, J. (2023). The effect of mutational robustness on the evolvability of multicellular organisms and eukaryotic cells. Journal of Evolutionary Biology 36:906–924. DOI: 10.1111/jeb.14180. https://doi.org/10.1111/jeb.14180
  14. Metzger, B. P. H. (2024). Genetic code robustness and protein evolvability are correlated and protein-specific. PLOS Biology 22:e3002627. DOI: 10.1371/journal.pbio.3002627. https://doi.org/10.1371/journal.pbio.3002627
  15. Kumawat, B., Lalejini, A., Acosta, M. M. & Zaman, L. (2024). Evolution takes multiple paths to evolvability when facing environmental change. Proceedings of the National Academy of Sciences USA 122:e2413930121. DOI: 10.1073/pnas.2413930121. https://doi.org/10.1073/pnas.2413930121
  16. Kharma, N. & Bédard-Couture, R. (2024). Robustness and evolvability: Revisited, redefined and applied. BioSystems 246:105281. DOI: 10.1016/j.biosystems.2024.105281. https://doi.org/10.1016/j.biosystems.2024.105281
  17. Hardy, N. B. (2025). How does robustness affect evolvability? Evolution 79:1996–2006. DOI: 10.1093/evolut/qpaf116. https://doi.org/10.1093/evolut/qpaf116
  18. Sappington, A. & Mohanty, V. (2025). Probabilistic genotype–phenotype maps reveal mutational robustness of RNA folding, spin glasses, and quantum circuits. Physical Review Research 7:013118. DOI: 10.1103/PhysRevResearch.7.013118. https://doi.org/10.1103/PhysRevResearch.7.013118
  19. Hulst, R. I., Govers, S. K., Michiels, J., Verstrepen, K. J. & van den Berg, P. (2025). The evolutionary role of mutational robustness: theoretical insights. Trends in Genetics (online 2025; volume 42, issue 2, pages 150–162, issue dated February 2026). DOI: 10.1016/j.tig.2025.10.011. https://doi.org/10.1016/j.tig.2025.10.011
  20. Hardy, N. B. (2026). The robustness and evolvability of continuously varying traits. Evolution 80:1532–1541. DOI: 10.1093/evolut/qpag082. https://doi.org/10.1093/evolut/qpag082
  21. Chakraborty, D., Rengaswamy, R. & Raman, K. (2026). UNified FramewOrk for reguLatory Dynamics (UNFOLD): Dissecting robustness, plasticity, evolvability and canalisation of biological function. PLOS Computational Biology 22:e1014289. DOI: 10.1371/journal.pcbi.1014289. https://doi.org/10.1371/journal.pcbi.1014289