[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-118407-en":3,"doc-seo-118407-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},118407,1099514067438,"River Wang","https://ap-avatar.wpscdn.com/avatar/100002539ee87300030?x-image-process=image/resize,m_fixed,w_180,h_180&k=1780474512215547542",8,"Research & Report","Machine learning in biological physics - From biomolecular prediction to design","The perspective argues that biological physics benefits most from combining theoretical modeling with machine learning rather than replacing one with the other. It links early neural-network ideas such as Hopfield networks to modern approaches including Potts models, Boltzmann machines, and transformer architectures through a shared energy representation. The text surveys how these integrated formulations address biomolecular structure, dynamics, function, evolution, and design, including protein structure prediction, improved molecular dynamics accuracy, and mutation-effect inference for evolutionary modeling. It further discusses protein design where learnable physical-model principles enable generation of sequences folding to target structures.","Downloaded from [https://www.pnas.org by \"RICE UNIVERSITY, FONDREN LIBRARY/MS 235\" on August 7, 2024 from IP address](https://www.pnas.org by \"RICE UNIVERSITY, FONDREN LIBRARY/MS 235\" on August 7, 2024 from IP address) [128.42.224.193](128.42.224.193).  \nPERSPECTIVE  \n OPEN ACCESS  \nMachine learning in biological physics: From biomolecular prediction to design  \nJonathan Martina,1, Marcos Lequerica Mateosb,1 ID , José N. Onuchicc,d,e,f,2, Ivan Coluzzab,g,2 , and Faruck Morcosa,h,2 Edited by Yuhai Tu, International Business Machines Corp., Yorktown Heights, NY; received September 4, 2023; accepted December 8, 2023, by Editorial Board Member Herbert Levine  \nMachine learning has been proposed as an alternative to theoretical modeling when dealing with complex problems in biological physics. However, in this perspective, we argue that a more successful approach is a proper combination of these two methodologies. We discuss how ideas coming from physical modeling neuronal processing led to early formulations of computational neural networks, e.g., Hopfield networks. We then show how modern learning approaches like Potts models, Boltzmann machines, and the transformer architecture are related to each other, specifically, through a shared energy representation. We summarize recent efforts to establish these connections and provide examples on how each of these formulations integrating physical modeling and machine learning have been successful in tackling recent problems in biomolecular structure, dynamics, function, evolution, and design. Instances include protein structure prediction; improvement in computational complexity and accuracy of molecular dynamics simulations; better inference of the effects of mutations in proteins leading to improved evolutionary modeling and finally how machine learning is revolutionizing protein engineering and design. Going beyond naturally existing protein sequences, a connection to protein design is discussed where synthetic sequences are able to fold to naturally occurring motifs driven by a model rooted in physical principles. We show that this model is “learnable” and propose its future use in the generation of unique sequences that can fold into a target structure.  \nPotts model j protein design j protein structure and dynamics j protein evolution j transformer model  \nIn recent years, a large number of fields of science have been impacted by theoretical and technical breakthroughsin the field of machine learning. These developmentsand prediction ability are fueled by the emergence of hardware that is optimized for learning architectures and the availability and storage of large amounts of high-quality data resulting from experimental efforts. Physics, and more specifically, biological physics is not an exception. On the contrary, biological physics is one of the sub-fields of science that has benefited the most by the convergence of large amounts of biological data and the development of modeling and learning approaches to unravel the mechanisms of biological phenomena. Clear examples include advancing our understanding of the sequence–structure–function relationships in biomolecules, the dynamics of protein folding, and biomedical applications. In this perspective, we aim to provide a glimpse at the state-of-the-art algorithms  \nin machine learning and how they are utilized for several applications in biological physics. We provide a nonexhaustive, but focused, account on how important modern learning algorithms, like the transformer architecture, are inherently connected to early developments in biological physics such as the Hopfield Network. We show how a mathematical representation of several learning algorithms in terms of “energy” functions unifies these formulationsand has been used for different applications and problems concerning biological phenomena. We focus on the study of biomolecules, their structures, functions, and dynamics. We also look into the problem of protein design and h","cbCaimSX3zFNh0wX","https://ap.wps.com/l/cbCaimSX3zFNh0wX","pdf",1923840,1,10,"English","en",105,"# Overview\n## Integrating modeling and learning in biological physics\n## Energy representations linking learning algorithms\n# Applications in biomolecular science\n## Protein structure, dynamics, and function\n## Protein evolution and mutation effects\n## Protein engineering and design\n# Future directions\n## Learnable energy Hamiltonian for target-structure generation","[{\"question\":\"Why does the perspective emphasize combining theoretical modeling and machine learning?\",\"answer\":\"It states that complex problems in biological physics are better addressed by a proper combination of both methodologies, not by treating machine learning as a standalone replacement for theory.\"},{\"question\":\"How are Hopfield networks and transformer architectures related in this work?\",\"answer\":\"The text explains that multiple modern learning approaches can be connected through a shared energy representation, tying transformers to earlier neural-network formulations such as Hopfield networks.\"},{\"question\":\"What is meant by the energy Hamiltonian being “learnable” in protein design?\",\"answer\":\"The perspective describes a model used to design proteins as learnable in a manner analogous to how evolutionary data helps infer amino-acid interactions within protein families.\"}]","Machine learning in biological physics - From biomolecular prediction to design | PDF",1785683474,25,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"machine-learning-in-biological-physics-from-biomolecular-prediction-to-design","",{"@graph":36,"@context":85},[37,54,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/machine-learning-in-biological-physics-from-biomolecular-prediction-to-design/118407/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-02",true,{"@type":65,"interactionType":66,"userInteractionCount":4},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"Why does the perspective emphasize combining theoretical modeling and machine learning?","Question",{"text":75,"@type":76},"It states that complex problems in biological physics are better addressed by a proper combination of both methodologies, not by treating machine learning as a standalone replacement for theory.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How are Hopfield networks and transformer architectures related in this work?",{"text":80,"@type":76},"The text explains that multiple modern learning approaches can be connected through a shared energy representation, tying transformers to earlier neural-network formulations such as Hopfield networks.",{"name":82,"@type":73,"acceptedAnswer":83},"What is meant by the energy Hamiltonian being “learnable” in protein design?",{"text":84,"@type":76},"The perspective describes a model used to design proteins as learnable in a manner analogous to how evolutionary data helps infer amino-acid interactions within protein families.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,134],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":21,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":21,"slug":133},"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]