[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-127397-en":3,"doc-seo-127397-105":31,"detail-sidebar-cat-0-en-105":92},{"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":20,"is_deleted":4,"is_public":21,"is_downloadable":21,"audit_status":21,"page_count":22,"language":23,"language_code":24,"site_id":25,"html_lang":24,"table_of_contents":26,"faqs":27,"seo_title":28,"seo_description":14,"update_tm":29,"read_time":30},127397,962085571259,"Theodora","https://ap-avatar.wpscdn.com/davatar_29158cc5080c5b710cf443261637dec0",8,"Research & Report","Recent advances in interpretable machine learning using structure-based protein representations","Recent advancements in machine learning are accelerating structural biology, especially through AlphaFold and related systems for protein structure prediction. This survey presents representation methods for protein 3D structures across resolutions and highlights interpretable ML approaches that support predicting protein structures, protein functions, and protein-protein interactions. Emphasis is placed on interpreting and visualizing model inference for structure-based representations to enable knowledge discovery. Such interpretable methods can improve candidate screening, enhance design decisions, and further accelerate drug development and protein design workflows.","arXiv :2409 . 17726v1 [ cs .LG] 26 Sep 2024  \nRecent advances in interpretable machine learning using structure-based protein representations  \nLUIZ FELIPE VECCHIETTI, Institute for Basic Science, South Korea  \nMINJI LEE, Korea Advanced Institute of Science and Technology, South Korea  \nBEGENCH HANGELDIYEV, Korea Advanced Institute of Science and Technology, South Korea HYUNKYU JUNG, Korea Advanced Institute of Science and Technology, South Korea HAHNBEOM PARK, Korea Institute of Science and Technology, South Korea  \nTAE-KYUN KIM, Korea Advanced Institute of Science and Technology, South Korea MEEYOUNG CHA, Max Planck Institute for Security and Privacy, Germany  \nHO MIN KIM, Korea Advanced Institute of Science and Technology, South Korea  \nRecent advancements in machine learning (ML) are transforming the field of structural biology. For example, AlphaFold, a groundbreaking neural network for protein structure prediction, has been widely adopted by researchers. The availability of easy-to-use interfaces and interpretable outcomes from the neural network architecture, such as the confidence scores used to color the predicted structures, have made AlphaFold accessible even to non-ML experts. In this paper, we present various methods for representing protein 3D structures from low-to high-resolution, and show how interpretable ML methods can support tasks such as predicting protein structures, protein function, and protein-protein interactions. This survey also emphasizes the significance of interpreting and visualizing ML-based inference for structure-based protein representations that enhance interpretability and knowledge discovery. Developing such interpretable approaches promises to further accelerate fields including drug development and protein design.  \nAdditional Key Words and Phrases: machine learning, artificial intelligence, interpretability, structural biology, protein functionality prediction, protein-protein interactions, protein structure prediction  \n1 INTRODUCTION  \nRecent advances in machine learning (ML) models have transformed the field of protein science, especially structural biology. Proteins are a building block of life, and their function is closely related to their three-dimensional (3D) structure. Indeed, understanding protein structure can often elucidate protein function. However, predicting how a protein folds from its amino acid sequence, which is represented as a list of text characters, has been a major challenge in biology for several decades. Recently, AlphaFold2 (AF2) [43], a groundbreaking neural network architecture, has made remarkable progress in accurately predicting protein structures that closely resemble experimental data. The rapid and widespread adoption of AF2 by the research community, even those lacking a theoretical background in ML, has been facilitated by user-friendly interfaces like ColabFold [58] . Moreover, interpretable modules in AF2’s neural network architecture, which are directly visualized within protein structures, help researchers interpret the ML model predictions and even apply the system to tasks such as de novo protein design. In particular, the predicted local distance difference test (pLDDT) and predicted aligned error (pAE) assist in screening candidates when designing novel protein binders. These  \nAuthors’ addresses: Luiz Felipe Vecchietti, [lfelipesv@ibs.re.kr](lfelipesv@ibs.re.kr), Institute for Basic Science, Daejeon, South Korea; Minji Lee, Korea Advanced Institute of Science and Technology, Daejeon, South Korea, [haewon_lee@kaist.ac.kr](haewon_lee@kaist.ac.kr); Begench Hangeldiyev, Korea Advanced Institute of Science and Technology, Daejeon, South Korea, [begahan@kaist.ac.kr](begahan@kaist.ac.kr); Hyunkyu Jung, Korea Advanced Institute of Science and Technology, Daejeon, South Korea, dino8egg@kaist. [ac.kr](ac.kr); Hahnbeom Park, Korea Institute of Science and Technology, Seoul, South Korea, [hahnbeom@kist.re.kr](hahnbeom@kist.re.kr); Tae-Kyun Kim, Kore","cbCaijZrp6vRD7fH","https://ap.wps.com/l/cbCaijZrp6vRD7fH","pdf",11171394,2,1,19,"English","en",105,"# Introduction\n## Interpretable machine learning and explainability approaches\n## Role of protein structure representations\n## AlphaFold2 metrics for candidate prioritization","[{\"question\":\"What does the survey focus on in interpretable machine learning for structural biology?\",\"answer\":\"It focuses on interpretable ML methods that use structure-based protein representations to support tasks such as protein structure prediction, protein function prediction, and protein-protein interaction inference.\"},{\"question\":\"How does AlphaFold2 enable interpretation of its predictions?\",\"answer\":\"AlphaFold2 provides interpretable, structure-linked outputs such as confidence scores (e.g., pLDDT) and predicted aligned error (pAE), which help screen and prioritize candidates.\"},{\"question\":\"What two broad categories of explanation methods are described for ML models?\",\"answer\":\"The survey distinguishes post-hoc methods that explain predictions for existing models and inherently explainable models that produce both the objective metric and human-understandable explanations.\"}]","Recent advances in interpretable machine learning using structure-based protein representations | PDF",1785938679,48,{"code":4,"msg":32,"data":33},"ok",{"site_id":25,"language":24,"slug":34,"title":13,"keywords":35,"description":14,"schema_data":36,"social_meta":87,"head_meta":89,"extra_data":91,"updated_unix":29},"recent-advances-in-interpretable-machine-learning-using-structure-based-protein-representations","",{"@graph":37,"@context":86},[38,54,69],{"@type":39,"itemListElement":40},"BreadcrumbList",[41,45,48,51],{"item":42,"name":43,"@type":44,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":46,"name":47,"@type":44,"position":20},"https://docshare.wps.com/document/","Document",{"item":49,"name":12,"@type":44,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":44,"position":53},"https://docshare.wps.com/document/recent-advances-in-interpretable-machine-learning-using-structure-based-protein-representations/127397/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":24,"description":14,"dateModified":62,"datePublished":63,"encodingFormat":61,"isAccessibleForFree":64,"interactionStatistic":65},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":42,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-22","2026-08-05",true,{"@type":66,"interactionType":67,"userInteractionCount":20},"InteractionCounter",{"@type":68},"ViewAction",{"@type":70,"mainEntity":71},"FAQPage",[72,78,82],{"name":73,"@type":74,"acceptedAnswer":75},"What does the survey focus on in interpretable machine learning for structural biology?","Question",{"text":76,"@type":77},"It focuses on interpretable ML methods that use structure-based protein representations to support tasks such as protein structure prediction, protein function prediction, and protein-protein interaction inference.","Answer",{"name":79,"@type":74,"acceptedAnswer":80},"How does AlphaFold2 enable interpretation of its predictions?",{"text":81,"@type":77},"AlphaFold2 provides interpretable, structure-linked outputs such as confidence scores (e.g., pLDDT) and predicted aligned error (pAE), which help screen and prioritize candidates.",{"name":83,"@type":74,"acceptedAnswer":84},"What two broad categories of explanation methods are described for ML models?",{"text":85,"@type":77},"The survey distinguishes post-hoc methods that explain predictions for existing models and inherently explainable models that produce both the objective metric and human-understandable explanations.","https://schema.org",{"og:url":52,"og:type":88,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":90,"canonical":52},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":93},[94,98,102,106,111,116,121,124,129,132,136],{"id":21,"doc_module":4,"doc_module_name":47,"category_name":95,"show_sort_weight":96,"slug":97},"Story & Novel",90,"story-novel",{"id":20,"doc_module":4,"doc_module_name":47,"category_name":99,"show_sort_weight":100,"slug":101},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":47,"category_name":103,"show_sort_weight":104,"slug":105},"Exam",70,"exam",{"id":107,"doc_module":4,"doc_module_name":47,"category_name":108,"show_sort_weight":109,"slug":110},5,"Comic",60,"comic",{"id":112,"doc_module":4,"doc_module_name":47,"category_name":113,"show_sort_weight":114,"slug":115},6,"Technology",50,"technology",{"id":117,"doc_module":4,"doc_module_name":47,"category_name":118,"show_sort_weight":119,"slug":120},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":47,"category_name":12,"show_sort_weight":122,"slug":123},30,"research-report",{"id":125,"doc_module":4,"doc_module_name":47,"category_name":126,"show_sort_weight":127,"slug":128},9,"Religion & Spirituality",20,"religion-spirituality",{"id":127,"doc_module":4,"doc_module_name":47,"category_name":130,"show_sort_weight":127,"slug":131},"World Cup","world-cup",{"id":133,"doc_module":4,"doc_module_name":47,"category_name":134,"show_sort_weight":133,"slug":135},10,"Lifestyle","lifestyle",{"id":22,"doc_module":4,"doc_module_name":47,"category_name":137,"show_sort_weight":107,"slug":138},"General","general"]