Learning the Shape of Evolutionary Landscapes: Geometric Deep Learning Reveals Hidden Structure in Phenotype-to-Fitness Maps
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 driv...
Uložené v:
| Vydané v: | bioRxiv |
|---|---|
| Hlavní autori: | , , |
| Médium: | Paper |
| Jazyk: | English |
| Vydavateľské údaje: |
Cold Spring Harbor Laboratory
07.05.2025
|
| Vydanie: | 1.1 |
| Predmet: | |
| ISSN: | 2692-8205 |
| On-line prístup: | Získať plný text |
| Tagy: |
Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
|
| 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 (RHVAE) 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 E. 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. |
|---|---|
| AbstractList | 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 (RHVAE) 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 E. 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. |
| Author | Mani, Madhav Petrov, Dmitri A. Razo-Mejia, Manuel |
| Author_xml | – sequence: 1 givenname: Manuel orcidid: 0000-0002-9510-0527 surname: Razo-Mejia fullname: Razo-Mejia, Manuel organization: Department of Biology, Stanford University – sequence: 2 givenname: Madhav orcidid: 0000-0002-5812-4167 surname: Mani fullname: Mani, Madhav email: madhav.mani@gmail.com organization: Department of Engineering Sciences and Applied Mathematics, Northwestern University – sequence: 3 givenname: Dmitri A. orcidid: 0000-0002-3664-9130 surname: Petrov fullname: Petrov, Dmitri A. email: madhav.mani@gmail.com organization: Chan Zuckerberg Biohub |
| BookMark | eNo9kEFLwzAcxYMoOOc-gLccvXQmaZqm3mRuU6gobveSJv-6yJaUJBsO_PBWJsKDd3j8Hrx3hc6dd4DQDSVTSgm9Y4QVUzKonIqCCSrO0IiJimWSkeISTWL8JISwStC85CP0XYMKzroPnDaAVxvVA_Ydnh_8dp-sdyocca2ciXpI4j1egt9BClbjR4Ae_9PvcAC1jfjJGgMOr1LY67QPgK3DbxtwPh17yJLPFjY5iBG_qD5eo4tugGDy52O0XszXs6esfl0-zx7qrC2FyDSvONeatFXHuTKMdzRnALqTRhmuS9BcaQmGSyC8lLwjYCoNpJWSFjnofIxuT7Wt9eHLHpo-2N0wrPn9qiGDyub0Vf4DeztjdQ |
| ContentType | Paper |
| Copyright | 2025, Posted by Cold Spring Harbor Laboratory |
| Copyright_xml | – notice: 2025, Posted by Cold Spring Harbor Laboratory |
| DBID | FX. |
| DOI | 10.1101/2025.05.07.652616 |
| DatabaseName | bioRxiv |
| DatabaseTitleList | |
| DeliveryMethod | fulltext_linktorsrc |
| Discipline | Biology |
| EISSN | 2692-8205 |
| Edition | 1.1 |
| ExternalDocumentID | 2025.05.07.652616v1 |
| GroupedDBID | 8FE 8FH AFKRA ALMA_UNASSIGNED_HOLDINGS BBNVY BENPR BHPHI FX. HCIFZ LK8 M7P NQS PHGZT PIMPY PROAC RHI |
| ID | FETCH-LOGICAL-b766-c4944cc0b9f44ad24f132eecf8dad4c7ec4ac8ed48e04784f0ed9ce0b88153ec3 |
| IngestDate | Sat May 10 15:10:20 EDT 2025 |
| IsDoiOpenAccess | true |
| IsOpenAccess | true |
| IsPeerReviewed | false |
| IsScholarly | false |
| Language | English |
| License | This pre-print is available under a Creative Commons License (Attribution 4.0 International), CC BY 4.0, as described at http://creativecommons.org/licenses/by/4.0 |
| LinkModel | OpenURL |
| MergedId | FETCHMERGED-LOGICAL-b766-c4944cc0b9f44ad24f132eecf8dad4c7ec4ac8ed48e04784f0ed9ce0b88153ec3 |
| Notes | Competing Interest Statement: The authors have declared no competing interest. |
| ORCID | 0000-0002-9510-0527 0000-0002-5812-4167 0000-0002-3664-9130 |
| OpenAccessLink | https://www.biorxiv.org/content/10.1101/2025.05.07.652616 |
| PageCount | 22 |
| ParticipantIDs | biorxiv_primary_2025_05_07_652616 |
| PublicationCentury | 2000 |
| PublicationDate | 20250507 |
| PublicationDateYYYYMMDD | 2025-05-07 |
| PublicationDate_xml | – month: 5 year: 2025 text: 20250507 day: 7 |
| PublicationDecade | 2020 |
| PublicationTitle | bioRxiv |
| PublicationYear | 2025 |
| Publisher | Cold Spring Harbor Laboratory |
| Publisher_xml | – name: Cold Spring Harbor Laboratory |
| References | Arvanitidis, Hansen, Hauberg (2025.05.07.652616v1.19) 2018 Weinreich, Delaney, DePristo, Hartl (2025.05.07.652616v1.11) 2006 Wagner (2025.05.07.652616v1.12) 2008; 275 Kingma, Welling (2025.05.07.652616v1.18) 2014 Martin, Lequerica Mateos, Onuchic, Coluzza, Morcos (2025.05.07.652616v1.14) 2024; 121 Petti, Reddy, Desai (2025.05.07.652616v1.17) 2023 Ascensao, Wetmore, Good, Arkin, Hallatschek (2025.05.07.652616v1.10) 2023; 14 Łuksza, Lässig (2025.05.07.652616v1.3) 2014; 507 Lässig, Mustonen, Walczak (2025.05.07.652616v1.1) 2017; 1 Chadebec, Allassonnière (2025.05.07.652616v1.24) 2022 Ghosh, Kinsler, Good, Petrov (2025.05.07.652616v1.16) 2025 Orr (2025.05.07.652616v1.23) 2005; 6 Louis (2025.05.07.652616v1.5) 2016; 58 Iwasawa, Maeda, Shibai, Kotani, Kawada, Furusawa (2025.05.07.652616v1.9) 2022 Kaneko, Furusawa, Yomo (2025.05.07.652616v1.20) 2015; 5 Łuksza (2025.05.07.652616v1.4) 2017; 551 Hauberg (2025.05.07.652616v1.22) 2019 Chadebec, Mantoux, Allassonnière (2025.05.07.652616v1.15) 2020 Kinsler, Geiler-Samerotte, Petrov (2025.05.07.652616v1.7) 2020; 9 Furusawa, Kaneko (2025.05.07.652616v1.13) 2018; 97 Manrubia (2025.05.07.652616v1.2) 2021; 38 Shoval, Sheftel, Shinar, Hart, Ramote, Mayo, Dekel, Kavanagh, Alon (2025.05.07.652616v1.6) 2012; 336 Maeda, Iwasawa, Kotani, Sakata, Kawada, Horinouchi, Sakai, Tanabe, Furusawa (2025.05.07.652616v1.8) 2020; 11 Russo, Husain, Murugan (2025.05.07.652616v1.21) 2025 |
| References_xml | – start-page: e3001920 year: 2022 ident: 2025.05.07.652616v1.9 article-title: Analysis of the Evolution of Resistance to Multiple Antibiotics Enables Prediction of the Escherichia Coli Phenotype-Based Fitness Landscape publication-title: PLOS Biology – volume: 121 start-page: e2311807121 year: 2024 ident: 2025.05.07.652616v1.14 article-title: Machine Learning in Biological Physics: From Biomolecular Prediction to Design publication-title: Proceedings of the National Academy of Sciences – volume: 58 start-page: 107 year: 2016 end-page: 16 ident: 2025.05.07.652616v1.5 article-title: Contingency, Convergence and Hyper-Astronomical Numbers in Biological Evolution publication-title: Studies in History and Philosophy of Science Part C: Studies in History and Philosophy of Biological and Biomedical Sciences – volume: 275 start-page: 91 year: 2008 end-page: 100 ident: 2025.05.07.652616v1.12 article-title: Robustness and Evolvability: A Paradox Resolved publication-title: Proceedings of the Royal Society B: Biological Sciences – year: 2018 ident: 2025.05.07.652616v1.19 publication-title: Latent Space Oddity: On the Curvature of Deep Generative Models – volume: 551 start-page: 517 year: 2017 end-page: 20 ident: 2025.05.07.652616v1.4 article-title: A Neoantigen Fitness Model Predicts Tumour Response to Checkpoint Blockade Immunotherapy publication-title: Nature – volume: 9 start-page: 1 year: 2020 end-page: 52 ident: 2025.05.07.652616v1.7 article-title: Fitness Variation across Subtle Environ-mental Perturbations Reveals Local Modularity and Global Pleiotropy of Adaptation publication-title: eLife – year: 2025 ident: 2025.05.07.652616v1.16 article-title: Low-Dimensional Genotype-Fitness Mapping across Divergent Environments Suggests a Limiting Functions Model of Fitness publication-title: bioRxiv – volume: 14 start-page: 248 year: 2023 ident: 2025.05.07.652616v1.10 article-title: Quantifying the Local Adaptive Landscape of a Nascent Bacterial Community publication-title: Nature Communications – volume: 11 start-page: 5970 year: 2020 ident: 2025.05.07.652616v1.8 article-title: High-Throughput Laboratory Evolution Reveals Evolutionary Constraints in Escherichia Coli publication-title: Nature Communications – year: 2022 ident: 2025.05.07.652616v1.24 publication-title: A Geometric Perspective on Variational Autoencoders – volume: 507 start-page: 57 year: 2014 end-page: 61 ident: 2025.05.07.652616v1.3 article-title: A Predictive Fitness Model for Influenza publication-title: Nature – volume: 336 start-page: 1157 year: 2012 end-page: 60 ident: 2025.05.07.652616v1.6 article-title: Evolutionary Trade-Offs, Pareto Optimality, and the Geometry of Phenotype Space publication-title: Science – year: 2014 ident: 2025.05.07.652616v1.18 publication-title: Auto-Encoding Variational Bayes – volume: 5 start-page: 011014 year: 2015 ident: 2025.05.07.652616v1.20 article-title: Universal Relationship in Gene-Expression Changes for Cells in Steady-Growth State publication-title: Physical Review X – volume: 1 start-page: 0077 year: 2017 ident: 2025.05.07.652616v1.1 article-title: Predicting Evolution publication-title: Nature Ecology & Evolution – volume: 6 start-page: 119 year: 2005 end-page: 27 ident: 2025.05.07.652616v1.23 article-title: The Genetic Theory of Adaptation: A Brief History publication-title: Nature Reviews Genetics – volume: 97 start-page: 042410 year: 2018 ident: 2025.05.07.652616v1.13 article-title: Formation of Dominant Mode by Evolution in Biological Systems publication-title: Physical Review E – volume: 38 start-page: 55 year: 2021 end-page: 106 ident: 2025.05.07.652616v1.2 article-title: From Genotypes to Organisms: State-of-the-art and Perspectives of a Cornerstone in Evolutionary Dynamics publication-title: Physics of Life Reviews – year: 2023 ident: 2025.05.07.652616v1.17 article-title: Inferring Sparse Structure in Genotype-Phenotype Maps publication-title: Genetics – year: 2020 ident: 2025.05.07.652616v1.15 publication-title: Geometry-Aware Hamiltonian Variational Auto-Encoder – year: 2006 ident: 2025.05.07.652616v1.11 article-title: Darwinian Evolution Can Follow Only Very Few Mutational Paths to Fitter Proteins publication-title: Science – year: 2025 ident: 2025.05.07.652616v1.21 article-title: Soft Modes as a Predictive Framework for Low Dimensional Biological Systems across Scales publication-title: Annual Review of Biophysics – year: 2019 ident: 2025.05.07.652616v1.22 publication-title: Only Bayes Should Learn a Manifold (on the Estimation of Differential Geometric Structure from Data) |
| SSID | ssj0002961374 |
| Score | 1.7571771 |
| SecondaryResourceType | preprint |
| Snippet | Elucidating the complex relationships between genotypes, phenotypes, and fitness remains one of the fundamental challenges in evolutionary biology. Part of the... |
| SourceID | biorxiv |
| SourceType | Open Access Repository |
| SubjectTerms | Evolutionary Biology |
| Title | Learning the Shape of Evolutionary Landscapes: Geometric Deep Learning Reveals Hidden Structure in Phenotype-to-Fitness Maps |
| URI | https://www.biorxiv.org/content/10.1101/2025.05.07.652616 |
| hasFullText | 1 |
| inHoldings | 1 |
| isFullTextHit | |
| isPrint | |
| link | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwtV3da9swEBcj2WBv-2TfaLC34i5VVUva2zba9aEJIctD34Isn4jHapkkNenYH7872bXD2ofuYWCMEZY_7idOd6ef7hj7YG0mMushodkSHRRPJIBMJd4646ywqXY6FptQk4k-PzfTlpC5juUEVFnq7dZU_xVqbEOwaevsP8DdPRQb8BpBxzPCjuc7AX92Hewgk_L70lYNe7BuX0osuTPa30vMp8iH-wbhgupqOVQ-UO11_WdQAyVXPqUsI1TolzLN0npDgXpzCWWg8G2yCclJsYkKc2yr9a6xmxVhti3qfinpV0jG8KOh545teQk9KbEpLYWt-dLWvb7erEId1eJFgV-493l_N0ohjiInUPVkofATzecYqqTdSTi88VfjIA8t2xmizhOpQQUt4k7sW7R7rCpAj4_ZVtV-eoQO4F-ZtOPcfOOeGj3koaBVyAEbfjmeTGddJE4YNGmUbJe88R0fb_RG5whFtkKR7Rgf80dsOEWoVo_ZPSifsAdN9dCrp-z3NVIckeYRaR4830Wa90h_4h3OnHDmXe8WZ97gzDuceVHy23DmhPMzNj85nn89TdpiGkmm0jRx0kjp3CgzXkqbC-kPDgWA8zq3uXQKnLROQy41UL4m6UeQGwejTGucE8EdPmeDMpTwgnEjI18hS8EbCamy2qQKDqSVRmt011-y962wFlWTMWVBAl2M8FCLRqCv7nDPa_awH0lv2AD_Ht6y-67eFOvVuxbFP7HhYbg |
| linkProvider | ProQuest |
| openUrl | ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Learning+the+Shape+of+Evolutionary+Landscapes%3A+Geometric+Deep+Learning+Reveals+Hidden+Structure+in+Phenotype-to-Fitness+Maps&rft.jtitle=bioRxiv&rft.au=Razo-Mejia%2C+Manuel&rft.au=Mani%2C+Madhav&rft.au=Petrov%2C+Dmitri+A.&rft.date=2025-05-07&rft.pub=Cold+Spring+Harbor+Laboratory&rft.eissn=2692-8205&rft_id=info:doi/10.1101%2F2025.05.07.652616&rft.externalDocID=2025.05.07.652616v1 |