[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-639294-105":59,"doc-detail-639294-en":130},{"code":4,"msg":5,"data":6},0,"success",[7,13,18,23,28,33,38,43,48,51,55],{"id":8,"doc_module":4,"doc_module_name":9,"category_name":10,"show_sort_weight":11,"slug":12},1,"Document","Story & Novel",90,"story-novel",{"id":14,"doc_module":4,"doc_module_name":9,"category_name":15,"show_sort_weight":16,"slug":17},2,"Literature",80,"literature",{"id":19,"doc_module":4,"doc_module_name":9,"category_name":20,"show_sort_weight":21,"slug":22},4,"Exam",70,"exam",{"id":24,"doc_module":4,"doc_module_name":9,"category_name":25,"show_sort_weight":26,"slug":27},5,"Comic",60,"comic",{"id":29,"doc_module":4,"doc_module_name":9,"category_name":30,"show_sort_weight":31,"slug":32},6,"Technology",50,"technology",{"id":34,"doc_module":4,"doc_module_name":9,"category_name":35,"show_sort_weight":36,"slug":37},7,"Healthcare",40,"healthcare",{"id":39,"doc_module":4,"doc_module_name":9,"category_name":40,"show_sort_weight":41,"slug":42},8,"Research & Report",30,"research-report",{"id":44,"doc_module":4,"doc_module_name":9,"category_name":45,"show_sort_weight":46,"slug":47},9,"Religion & Spirituality",20,"religion-spirituality",{"id":46,"doc_module":4,"doc_module_name":9,"category_name":49,"show_sort_weight":46,"slug":50},"World Cup","world-cup",{"id":52,"doc_module":4,"doc_module_name":9,"category_name":53,"show_sort_weight":52,"slug":54},10,"Lifestyle","lifestyle",{"id":56,"doc_module":4,"doc_module_name":9,"category_name":57,"show_sort_weight":24,"slug":58},19,"General","general",{"code":4,"msg":60,"data":61},"ok",{"site_id":62,"language":63,"slug":64,"title":65,"keywords":66,"description":67,"schema_data":68,"social_meta":123,"head_meta":125,"extra_data":127,"updated_unix":129},105,"en","pring-rethinking-protein-protein-interaction-prediction-from-pairs-to-graphs-639294","PRING - Rethinking Protein-Protein Interaction Prediction from Pairs to Graphs","","Deep learning methods for protein-protein interaction (PPI) prediction show strong pairwise accuracy, yet existing benchmarks mainly evaluate isolated pairs and ignore whether models can reconstruct biologically meaningful PPI networks. PRING introduces a graph-level benchmark to assess network reconstruction, curating a high-quality multi-species dataset with 21,484 proteins and 186,818 interactions. It includes redundancy and leakage control, and two topology-oriented and three function-oriented evaluation paradigms. Experiments across four model categories highlight limitations in recovering structural and functional properties, guiding more real-world applications.",{"@graph":69,"@context":122},[70,84,105],{"@type":71,"itemListElement":72},"BreadcrumbList",[73,77,79,82],{"item":74,"name":75,"@type":76,"position":8},"https://docshare.wps.com","Home","ListItem",{"item":78,"name":9,"@type":76,"position":14},"https://docshare.wps.com/document/",{"item":80,"name":40,"@type":76,"position":81},"https://docshare.wps.com/document/research-report/",3,{"item":83,"name":65,"@type":76,"position":19},"https://docshare.wps.com/document/pring-rethinking-protein-protein-interaction-prediction-from-pairs-to-graphs-639294/639294/",{"url":83,"name":65,"@type":85,"image":86,"author":91,"headline":65,"publisher":94,"fileFormat":97,"inLanguage":63,"description":67,"dateModified":98,"datePublished":99,"encodingFormat":97,"isAccessibleForFree":100,"interactionStatistic":101},"DigitalDocument",{"url":87,"@type":88,"width":89,"height":90},"https://docshare.wps.com/thumbnails/pring-rethinking-protein-protein-interaction-prediction-from-pairs-to-graphs-639294/639294.png","ImageObject",300,407,{"name":92,"@type":93},"Elsa","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-10-31","2026-10-22",true,{"@type":102,"interactionType":103,"userInteractionCount":39},"InteractionCounter",{"@type":104},"ViewAction",{"@type":106,"mainEntity":107},"FAQPage",[108,114,118],{"name":109,"@type":110,"acceptedAnswer":111},"What problem does PRING address in PPI prediction benchmarks?","Question",{"text":112,"@type":113},"PRING targets the gap where existing benchmarks largely measure isolated pairwise predictions and do not evaluate whether models can reconstruct cohesive, biologically meaningful PPI networks.","Answer",{"name":115,"@type":110,"acceptedAnswer":116},"What does the PRING benchmark provide in terms of data and evaluations?",{"text":117,"@type":113},"PRING provides a multi-species PPI network dataset (21,484 proteins, 186,818 interactions) and defines complementary evaluation paradigms: topology-oriented tasks and function-oriented tasks.",{"name":119,"@type":110,"acceptedAnswer":120},"Which types of models are evaluated, and what do the experiments indicate?",{"text":121,"@type":113},"Experiments cover sequence similarity-based, naive sequence-based, protein language model-based, and structure-based approaches, showing limitations in recovering both structural and functional properties of PPI networks.","https://schema.org",{"og:url":83,"og:type":124,"og:title":65,"og:site_name":95,"og:description":67},"article",{"robots":126,"canonical":83},"index,follow",{"doc_id":128,"site_id":62},639294,1792886612,{"code":4,"msg":5,"data":131},{"doc_id":128,"user_id":132,"nickname":92,"user_avatar":133,"doc_module":4,"category_id":39,"category_name":40,"doc_title":65,"doc_description":67,"doc_content":134,"file_id":135,"file_url":136,"file_type":137,"file_size":138,"view_count":39,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":139,"language":140,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":141,"faqs":142,"seo_title":143,"seo_description":67,"update_tm":144,"read_time":145},137455077381,"https://ap-avatar.wpscdn.com/davatar_994ba38a5ba835b3df7d355c54d3ed8d","arXiv :2507 .05101v2 [ cs .LG] 22 Oct 2025  \nPRING: Rethinking Protein-Protein Interaction Prediction from Pairs to Graphs  \nXinzhe Zheng 1 ,2 ∗ Hao Du2 ∗ Fanding Xu 1 ,7 ∗ Jinzhe Li2 ,4 ∗ Zhiyuan Liu 1† Wenkang Wang 1 ,8 Tao Chen3 ,4 Wanli Ouyang2 ,6 ,9  \nStan Z. Li5 Yan Lu2 ,6† Nanqing Dong2 ,3† Yang Zhang 1†  \n1National University of Singapore 2 Shanghai Artificial Intelligence Laboratory  \n3 Shanghai Innovation Institute 4Fudan University 5Westlake University  \n6 Chinese University of Hong Kong 7Xi’an Jiaotong University  \n8 Central South University 9 Shenzhen Loop Area Institute  \nAbstract  \nDeep learning-based computational methods have achieved promising results in predicting protein-protein interactions (PPIs) . However, existing benchmarks predominantly focus on isolated pairwise evaluations, overlooking a model’s capability to reconstruct biologically meaningful PPI networks, which is crucial for biology research. To address this gap, we introduce PRING, the first comprehensive benchmark that evaluates PRotein-protein INteraction prediction from a Graph-level perspective. PRING curates a high-quality, multi-species PPI network dataset comprising 21,484 proteins and 186,818 interactions, with well-designed strategies to address both data redundancy and leakage. Building on this golden-standard dataset, we establish two complementary evaluation paradigms: (1) topologyoriented tasks, which assess intra and cross-species PPI network construction, and  \n(2) function-oriented tasks, including protein complex pathway prediction, GO module analysis, and essential protein justification. These evaluations not only reflect the model’s capability to understand the network topology but also facilitate protein function annotation, biological module detection, and even disease mechanism analysis. Extensive experiments on four representative model categories, consisting of sequence similarity-based, naive sequence-based, protein language model-based, and structure-based approaches, demonstrate that current PPI models have potential limitations in recovering both structural and functional properties of PPI networks, highlighting the gap in supporting real-world biological applications.  \nWe believe PRING provides a reliable platform to guide the development of more effective PPI prediction models for the community. The dataset and source code of PRING are available at [https://github.com/SophieSarceau/PRING](https://github.com/SophieSarceau/PRING).  \n1 Introduction  \nProtein-protein interactions (PPIs) refer to the physical or functional association between proteins, which is central to most biological processes, such as signal transduction [1], gene regulation [2], and immune response [3] . They also play a key role in the study of disease mechanisms [4] . For instance, the interactions between certain cellular proteins can trigger abnormal signaling pathways that promote tumor growth and metastasis in cancer [5] . By revealing those types of PPI, small molecules are developed to disrupt the interactions, thereby inducing cancer cell death [6] . Therefore, characterizing the organization of PPIs is fundamental for understanding the molecular basis of health and advancing drug discovery in precision medicine [7, 8] .  \n∗Equal contribution.  \n†Corresponding author.  \n39th Conference on Neural Information Processing Systems (NeurIPS 2025) Track on Datasets and Benchmarks.  \n|  |\n| --- |\n| d. Evaluation Metric |\n| \u003Cbr>PPI Network\u003Cbr>Graph Sim. Relative Density Degree Distribution Clustering Coefficient Spectral\u003Cbr>Essential Protein\u003Cbr>Precision@K Distribution Overlap\u003Cbr>Complex Pathway\u003Cbr>Pathway Precision Pathway Recall Pathway Connectivity\u003Cbr>GO Enrichment\u003Cbr>Functional Alignment Consistency Ratio (Jaccard Sim. ) |\n\n( , )  \n( , )  \n( , )  \n( , )  \n( , ) Protein Pairs  \nPPI Prediction Model  \nb. PRING Task Definition Topology-oriented Task  \nIntra-species  \nHUMAN  \nPPI Network  \nCross-species  \nARATH YEAST ECOLI  \nFunction-oriented Tas","cbCaimFlEFayIKHu","https://ap.wps.com/l/cbCaimFlEFayIKHu","pdf",3589991,38,"English","# Abstract\n# 1 Introduction\n## Protein-protein interactions and their biological significance\n## Experimental methods for PPI identification and limitations\n## Deep learning approaches and advances (AlphaFold, PLMs)\n## Motivation for graph-level, network reconstruction evaluation","[{\"question\":\"What problem does PRING address in PPI prediction benchmarks?\",\"answer\":\"PRING targets the gap where existing benchmarks largely measure isolated pairwise predictions and do not evaluate whether models can reconstruct cohesive, biologically meaningful PPI networks.\"},{\"question\":\"What does the PRING benchmark provide in terms of data and evaluations?\",\"answer\":\"PRING provides a multi-species PPI network dataset (21,484 proteins, 186,818 interactions) and defines complementary evaluation paradigms: topology-oriented tasks and function-oriented tasks.\"},{\"question\":\"Which types of models are evaluated, and what do the experiments indicate?\",\"answer\":\"Experiments cover sequence similarity-based, naive sequence-based, protein language model-based, and structure-based approaches, showing limitations in recovering both structural and functional properties of PPI networks.\"}]","PRING - Rethinking Protein-Protein Interaction Prediction from Pairs to Graphs | PDF",1792702876,96]