- Open Access
Learning the Shape of Evolutionary Landscapes: Geometric Deep Learning Reveals Hidden Structure in Phenotype-to-Fitness Maps
PRX Life 4, 013030 – Published 6 March, 2026
DOI: https://doi.org/10.1103/zyj4-qt72
Abstract
Elucidating the complex relationships between genotypes, phenotypes, and fitness remains one of the fundamental challenges in evolutionary biology. Part of the difficulty arises from the enormous number of possible genotypes and the lack of understanding of the underlying phenotypic differences driving adaptation. Here we present a computational method that takes advantage of modern high-throughput fitness measurements to learn a map from high-dimensional fitness profiles to a low-dimensional latent space in a geometry-informed manner. We demonstrate that our approach using a Riemannian Hamiltonian variational autoencoder outperforms traditional linear dimensionality reduction techniques by capturing the nonlinear structure of the phenotype-fitness map. When applied to simulated adaptive dynamics, we show that the learned latent space retains information about the underlying adaptive phenotypic space and accurately reconstructs complex fitness landscapes. We then apply this method to a dataset of high-throughput fitness measurements of Escherichia coli under different antibiotic pressures and demonstrate superior predictive power for out-of-sample data compared to linear approaches. Our work provides a data-driven implementation of Fisher's geometric model of adaptation, transforming it from a theoretical framework into an empirically grounded approach for understanding evolutionary dynamics using modern deep learning methods.
Physics Subject Headings (PhySH)
Collections
This article appears in the following collection:
PRX Life Collection on adaptation and learning
PRX Life presents a special collection on ‘Biological systems that learn’, curated by the organizers of the National Institute for Theory and Mathematics in Biology (NITMB) workshop together with early-career researchers. The papers in this collection illustrate how living systems acquire functions, showing it is useful to describe them as tunable matter with many individually adjustable degrees of freedom characterizing microscopic properties. Drawing inspiration from AI and non-biological systems that can acquire desired functions, the featured research highlights the timeliness of the topic and its potential to drive unifying frameworks for emergent collective function in living tunable matter.
Every article in this collection is subject to a rigorous peer review process, upholding the high standards applied to all papers in the journal. The PRX Life editorial team manages peer review and makes all editorial decisions.
Article Text
Supplemental Material
References (27)
- M. Lässig, V. Mustonen, and A. M. Walczak, Predicting evolution, Nat. Ecol. Evol. 1, 0077 (2017).
- S. Manrubia, J. A. Cuesta, J. Aguirre, S. E. Ahnert, L. Altenberg, A. V. Cano, P. Catalán, R. Diaz-Uriarte, S. F. Elena, J. A. García-Martín, P. Hogeweg, B. S. Khatri, J. Krug, A. A. Louis, N. S. Martin, J. L. Payne, M. J. Tarnowski, and M. Weiß, From genotypes to organisms: State-of-the-art and perspectives of a cornerstone in evolutionary dynamics, Phys. Life Rev. 38, 55 (2021).
- M. Łuksza and M. Lässig, A predictive fitness model for influenza, Nature (London) 507, 57 (2014).
- M. Łuksza, N. Riaz, V. Makarov, V. P. Balachandran, M. D. Hellmann, A. Solovyov, N. A. Rizvi, T. Merghoub, A. J. Levine, T. A. Chan, J. D. Wolchok, and B. D. Greenbaum, A neoantigen fitness model predicts tumour response to checkpoint blockade immunotherapy, Nature (London) 551, 517 (2017).
- A. A. Louis, Contingency, convergence and hyper-astronomical numbers in biological evolution, Stud. Hist. Philos. Sci. Part C: Stud. Hist. Philos. Biol. Biomed. Sci 58, 107 (2016).
- O. Shoval, H. Sheftel, G. Shinar, Y. Hart, O. Ramote, A. Mayo, E. Dekel, K. Kavanagh, and U. Alon, Evolutionary trade-offs, Pareto optimality, and the geometry of phenotype space, Science 336, 1157 (2012).
- J. A. Ascensao, K. M. Wetmore, B. H. Good, A. P. Arkin, and O. Hallatschek, Quantifying the local adaptive landscape of a nascent bacterial community, Nat. Commun. 14, 248 (2023).
- J. Iwasawa, T. Maeda, A. Shibai, H. Kotani, M. Kawada, and C. Furusawa, Analysis of the evolution of resistance to multiple antibiotics enables prediction of the Escherichia coli phenotype-based fitness landscape, PLoS Biol. 20, e3001920 (2022).
- G. Kinsler, K. Geiler-Samerotte, and D. A. Petrov, Fitness variation across subtle environmental perturbations reveals local modularity and global pleiotropy of adaptation, eLife 9, e61271 (2020).
- T. Maeda, J. Iwasawa, H. Kotani, N. Sakata, M. Kawada, T. Horinouchi, A. Sakai, K. Tanabe, and C. Furusawa, High-throughput laboratory evolution reveals evolutionary constraints in Escherichia coli, Nat. Commun. 11, 5970 (2020).
- C. Furusawa and K. Kaneko, Formation of dominant mode by evolution in biological systems, Phys. Rev. E 97, 042410 (2018).
- A. Wagner, Robustness and evolvability: A paradox resolved, Proc. R. Soc. B 275, 91 (2008).
- D. M. Weinreich, N. F. Delaney, M. A. DePristo, and D. L. Hartl, Darwinian evolution can follow only very few mutational paths to fitter proteins, Science 312, 111 (2006).
- J. Martin, M. L. Mateos, J. N. Onuchic, I. Coluzza, and F. Morcos, Machine learning in biological physics: From biomolecular prediction to design, Proc. Natl. Acad. Sci. USA 121, e2311807121 (2024).
- C. Chadebec, C. Mantoux, and S. Allassonnière, Geometry-aware Hamiltonian variational auto-encoder, arXiv:2010.11518.
- K. Kaneko, C. Furusawa, and T. Yomo, Universal relationship in gene-expression changes for cells in steady-growth state, Phys. Rev. X 5, 011014 (2015).
- I. Fragata, A. Blanckaert, M. A. D. Louro, D. A. Liberles, and C. Bank, Evolution in the light of fitness landscape theory, Trends Ecol. Evol. 34, 69 (2019).
- See Supplemental Material at http://link.aps.org/supplemental/10.1103/zyj4-qt72 for detailed derivations of the Metropolis-Hastings evolutionary dynamics, RHVAE architecture details, analysis of nonlinear manifold embeddings, cross-validation methodology, geodesic analysis, fitness landscape reconstruction, neural network architecture, latent space alignment, and Metropolis-Kimura evolutionary dynamics.
- O. M. Ghosh, G. Kinsler, B. H. Good, and D. A. Petrov, Low-dimensional genotype-fitness mapping across divergent environments suggests a limiting functions model of fitness, bioRxiv (2025), https://doi.org/10.1101/2025.04.05.647371.
- S. Petti, G. Reddy, and M. M. Desai, Inferring sparse structure in genotype-phenotype maps, Genetics 225, iyad127 (2023).
- D. P. Kingma and M. Welling, Auto-encoding variational Bayes, 2nd International Conference on Learning Representations, Conference Track Proceedings (ICLR, Canada, 2014).
- G. Arvanitidis, L. K. Hansen, and S. Hauberg, Latent space oddity: On the curvature of deep generative models, in Proceedings of 6th International Conference on Learning Representations, Conference Track Proceedings (ICLR, Vancouver, Canada, 2018).
- C. J. Russo, K. Husain, and A. Murugan, Soft modes as a predictive framework for low dimensional biological systems across scales, Annu. Rev. Biophys. 54, 401 (2025).
- S. Hauberg, Only Bayes should learn a manifold (on the estimation of differential geometric structure from data), arXiv:1806.04994.
- H. A. Orr, The genetic theory of adaptation: A brief history, Nat. Rev. Genet. 6, 119 (2005).
- C. Chadebec and S. Allassonnière, A geometric perspective on variational autoencoders, in Proceedings of the 36th International Conference on Neural Information Processing Systems, NIPS '22 (Curran Associates Inc., Red Hook, NY, USA, 2022), Vol. 35, pp. 19618–19630.
- M. Razo-Mejia, M. Mani, and D. A. Petrov, Code and data for “learning the shape of evolutionary landscapes: Geometric deep learning reveals hidden structure in phenotype-to-fitness maps,” bioRxiv (2026), https://doi.org/10.1101/2025.05.07.652616; GitHub repository, https://github.com/mrazomej/antibiotic_landscape.
