Evolution and information content of optimal gene regulatory architectures

Nick Barton, Gašper TkačikbioRxiv 2025 · DOI

TL;DR

This paper proposes that gene regulatory architectures evolve to optimize a "replicated genotype-phenotype map," where regulatory parameters are selected to maximize the number of genotypes associated with fit phenotypes. This principle, drawing from population genetics and information theory, suggests that evolved regulatory systems are akin to efficient codes, matching the frequency of phenotypes to their required representation across the genome.

What Problem Does This Paper Address?

The authors tackle the fundamental question of how gene regulatory architectures, particularly the sequences that control gene expression (like promoters and enhancers), evolve. While we know these sequences are subject to selection, the underlying "genotype-phenotype map" (how a DNA sequence translates to gene expression) itself can change due to evolving molecular machinery (e.g., transcription factors). This raises a crucial question: what kind of regulatory architectures should we expect to emerge under such evolutionary pressures, especially when a single change can affect many genes simultaneously? Existing models often treat individual regulatory elements in isolation, but this paper aims to understand the co-evolution of regulatory parameters that govern the entire network.

Methods

The core of the paper employs population genetics theory, specifically focusing on equilibrium distributions and the concept of "free fitness." They build an analogy between population genetics and statistical physics, using concepts like Boltzmann distributions and entropy. A key innovation is the development of a "replicated genotype-phenotype map" framework, where a common set of regulatory parameters influences the phenotype of multiple regulatory elements (like cis-regulatory elements, or CREs). They then use information theory, particularly the concept of coding efficiency, to derive optimality principles for these regulatory parameters. They illustrate their theory with a simplified "binary model" of transcription factor (TF) binding and then extend it to consider more complex scenarios like multiple TFs and probabilistic binding.

Datasets

Not discussed in the paper. The paper focuses on theoretical modeling and simulations rather than empirical datasets.

Key Insights and Ideas

  • "Replicated GP Map" is Key: The authors introduce the concept of a "replicated genotype-phenotype map," where a single set of regulatory parameters (e.g., TF concentration, binding site characteristics) governs the phenotype of many regulatory elements across the genome. This shared architecture is central to their evolutionary argument.
  • Optimality Principle: Maximize Fit Genotypes: Evolved regulatory parameters will favor configurations where fit phenotypes are associated with a large number of possible genotypes. This is a novel optimality principle derived from population genetics.
  • "Frequency Matching" as Optimal Coding: In the ideal scenario, regulatory architectures evolve to match the frequency of phenotypes required by selection to the fraction of sequences that can produce those phenotypes. This is analogous to optimal coding in information theory, where rare messages require more "code" (information).
  • Selection Drives Architecture, Not Just Sequences: The theory suggests that selection acts not only on individual regulatory sequences but also on the underlying parameters that define the entire regulatory architecture. This can happen even with weak selection on individual elements, provided there are many such elements.
  • Biophysical Constraints and "Overspecification": Real-world biophysical constraints (e.g., the continuous nature of TF binding probabilities) can prevent perfect "frequency matching," leading to "overspecification" – more information is needed than theoretically ideal to achieve the desired outcome.
  • Mutational Robustness Matters: In high mutation rate regimes, regulatory architectures evolve to be robust against mutations, favoring genotypes with many fit neighbors, which can also increase information requirements.

Key Results

  • The authors derive an optimality principle for regulatory parameters that maximizes free fitness, which balances expected log-fitness with genetic information.
  • In simplified models, they show that the optimal regulatory parameter (e.g., the fraction of sequences binding a TF) evolves to match the fraction of CREs under selection for binding (q* = f).
  • They demonstrate that this "frequency matching" principle is analogous to optimal coding in information theory, where the minimal information required is related to the entropy of the required phenotype frequencies.
  • The introduction of biophysical constraints (probabilistic binding) leads to a need for more information than theoretically optimal, a phenomenon termed "overspecification."
  • Strong mutation rates also increase information requirements by favoring mutational robustness.

Limitations and Future Work

  • Stationary Distributions: The theory primarily relies on equilibrium distributions, which may not always be reached in real biological systems, especially with rapid evolutionary changes or complex dynamics.
  • Mutation Types: The model's treatment of mutations is simplified, primarily focusing on point mutations and excluding structural changes like duplications or deletions.
  • Knowledge of Selection: The framework requires knowing the selective pressures (i.e., the required regulatory phenotypes). In practice, inferring these requirements from observed data is a significant challenge.
  • Future Work: The authors suggest extending the theory to include non-equilibrium dynamics, a wider range of mutation types, and developing methods for inferring regulatory requirements from genomic data. They also propose testing their predictions through evolutionary simulations and comparative genomics studies.

Summary updated

Abstract

Abstract Evolution of gene regulatory sequences is shaped by their genotype-phenotype (GP) map, but this map itself can evolve via changes in the molecular machinery that reads out the regulatory sequences. Using population genetics theory, we derive an optimality principle, implying that the evolved GP maps will associate fit phenotypes with larger numbers of possible genotypes. Mathematically, this builds on analogies between population genetics and statistical physics, as well as optimal coding in information theory. The results are particularly interesting in the context of transcriptional regulation, where optimal values of certain parameters that determine the gene regulatory architecture can be derived even without a complete knowledge of the molecular mechanisms involved. We illustrate the theory using a simplified model of transcription factor (TF) binding to cis-regulatory elements (CREs): the fraction of possible CREs across the entire genotype space that bind a given TF evolves to match the fraction of CREs under selection to bind that TF. Similar “frequency matching” results are expected to emerge generically, because selection implicitly optimizes the entire regulatory architecture that determines the “replicated GP map” for all CREs, while it simultaneously adapts towards individually fit CREs on that GP map. We discuss the subtleties and limitations of the theory related to biophysical constraints and the need for mutational robustness. Lastly, our theory suggests a possible mathematical definition for evolvability with predictive power that can be tested in evolutionary simulations and, ultimately, in comparative genetics studies.
BibTeX
@misc{barton2025evolution,
  author = {Nick Barton and Gašper Tkačik},
  title = {{Evolution and information content of optimal gene regulatory architectures}},
  year = {2025},
  doi = {10.1101/2025.06.10.657849},
  howpublished = {bioRxiv},
}

Loading related content...

Figure