Abstract: Engineers routinely design systems to be modular and symmetric in order to increase robustness to perturbations and to facilitate alterations at a later date. Biological structures also frequently exhibit modularity and symmetry, but the origin of such trends is much less well understood. It can be tempting to assume – by analogy to engineering design – that symmetry and modularity arise from natural selection. But evolution, unlike engineers, cannot plan ahead, and so these traits must also afford some immediate selective advantage which is hard to reconcile with the breadth of systems where symmetry is observed. Here we introduce an alternative non-adaptive hypothesis based on an algorithmic picture of evolution. It suggests that symmetric structures preferentially arise not just due to natural selection, but also because they require less specific information to encode, and are therefore much more likely to appear as phenotypic variation through random mutations. Arguments from algorithmic information theory can formalise this intuition, leading to the prediction that many genotype-phenotype maps are exponentially biased towards phenotypes with low descriptional complexity. A preference for symmetry is a special case of this bias towards compressible descriptions. We test these predictions with extensive biological data, showing that that protein complexes, RNA secondary structures, and a model gene-regulatory network all exhibit the expected exponential bias towards simpler (and more symmetric) phenotypes. Lower descriptional complexity also correlates with higher mutational robustness, which may aid the evolution of complex modular assemblies of multiple components.
Introduction: The Puzzle of Biological Symmetry
Why is the natural world so full of symmetry and modularity? From the hexagonal patterns of honeycombs to the bilateral symmetry of the human body, these traits are ubiquitous. In engineering, modular and symmetric designs are chosen intentionally for their robustness and ease of adaptation. In biology, the standard assumption has been that natural selection "chooses" these designs because they offer a survival advantage.
However, evolution lacks foresight. For a symmetric trait to be selected, it must first appear as a mutation. This paper, published in PNAS, addresses a fundamental gap in evolutionary theory: Is the "menu" of mutations offered to natural selection biased? The authors propose that symmetric phenotypes prevail not necessarily because they are "better," but because they are exponentially more likely to appear as random mutations—a phenomenon termed "arrival of the frequent."
The Algorithmic Hypothesis
The researchers utilize Algorithmic Information Theory (AIT) to formalize the concept of simplicity. They define the complexity of a phenotype (a biological structure) using Kolmogorov Complexity ($K(p)$)—the length of the shortest "program" needed to describe it. Symmetric structures are inherently "compressible"; one only needs to describe a single subunit and the rules for repeating it.
Borrowing the Coding Theorem from AIT, the authors predict that in many input-output maps (like the map from DNA to a protein shape), the probability $P(p)$ of a phenotype appearing is exponentially biased toward those with low descriptional complexity: $P(p) \approx 2^{-aK(p)-b}$.
Evidence from Protein Complexes
To test this hypothesis, the authors analyzed a curated set of 34,287 protein complexes from the Protein Data Bank (PDB). Because true Kolmogorov complexity is uncomputable, they used the number of unique interface types as a proxy for complexity in protein quaternary structures.
As shown in Figure 2, there is a striking exponential bias. In complexes consisting of six subunits (6-mers), the top three simplest topologies account for 76.8% of all complexes found in the PDB. Panel B illustrates that as the complexity $\tilde{K}$ (number of interface types) increases, the frequency of that structure in nature drops precipitously. Panel C further demonstrates that highly symmetric groups (like C6 and D3) appear in evolution far more frequently than their "fraction of space" (the number of possible ways to build them) would suggest.
Computational Validation: The Polyomino Model
The authors further validated their theory using a "Polyomino" model, where 2D tiles self-assemble into shapes based on binding rules encoded in a bit-string genotype.
Figure 4 shows that the computational model mirrors the biological data perfectly. In 50,000 evolutionary runs, low-complexity, high-symmetry structures (such as those with D4 symmetry) were common, while high-complexity, low-symmetry structures were rare. This confirms that the bias toward symmetry is a default mathematical property of information encoding in genotype-phenotype maps.
Broader Implications and Conclusion
The study extended these findings to RNA secondary structures and Gene Regulatory Networks (GRNs). In both cases, naturally occurring structures were significantly more compressible and simpler than expected by chance. For instance, the wild-type behavior of the budding yeast cell-cycle network was found to be one of the simplest possible outputs the model could produce.
This research suggests that much of the symmetry we see in nature has a non-adaptive origin. While natural selection still acts as a filter, removing simple structures that are non-functional, the "arrival of the frequent" ensures that selection primarily encounters simple, symmetric designs. This algorithmic bias provides a powerful, universal explanation for the emergence of order in the living world.
Summary updated