{"id":32205,"date":"2026-07-22T11:19:25","date_gmt":"2026-07-22T09:19:25","guid":{"rendered":"https:\/\/sano.science\/?p=32205"},"modified":"2026-07-22T11:36:28","modified_gmt":"2026-07-22T09:36:28","slug":"bringing-clarity-to-how-machine-learning-models-see-molecules","status":"publish","type":"post","link":"https:\/\/sano.science\/bringing-clarity-to-how-machine-learning-models-see-molecules\/","title":{"rendered":"Bringing clarity to how machine learning models &#8220;see&#8221; molecules\u00a0"},"content":{"rendered":"\n<p class=\"eplus-wrapper wp-block-paragraph\">Our researchers Adam Su\u0142ek and Tomasz Ko\u015bcio\u0142ek, together with their co-authors, have published a new paper in the <em>Journal of Computational Science<\/em>, addressing one of the key challenges in molecular machine learning: evaluating atom-level explanations.<\/p>\n\n\n\n<div style=\"height:10px\" aria-hidden=\"true\" class=\"wp-block-spacer eplus-wrapper\"><\/div>\n\n\n\n<h2 id=\"h-benchmarking-atom-level-explainability-against-pharmacophore-computed-labels-in-molecular-machine-learning\" class=\"wp-block-heading eplus-wrapper\">\u201cBenchmarking atom-level explainability against pharmacophore-computed labels in molecular machine learning\u201d<\/h2>\n\n\n\n<p class=\"eplus-wrapper wp-block-paragraph\"><strong>Authors:<\/strong> Adam Su\u0142ek, Jakub Klimczak, Jakub Jo\u0144czyk, Tomasz Ko\u015bcio\u0142ek, Tomasz Danel and Barbara Pucelik<\/p>\n\n\n\n<div style=\"height:10px\" aria-hidden=\"true\" class=\"wp-block-spacer eplus-wrapper\"><\/div>\n\n\n\n<p class=\"eplus-wrapper wp-block-paragraph\">Machine learning models can predict molecular properties with impressive accuracy\u2014but do their explanations identify atoms that are genuinely relevant to molecular structure and function?<\/p>\n\n\n\n<div style=\"height:10px\" aria-hidden=\"true\" class=\"wp-block-spacer eplus-wrapper\"><\/div>\n\n\n\n<p class=\"eplus-wrapper wp-block-paragraph\">The study introduces a controlled benchmark based on pharmacophores and geometric rules describing the spatial arrangement of functional elements within molecules. At the core of the framework is PharmacoScore, a metric that quantifies how closely atom-level model explanations align with pharmacophore-derived reference annotations.<\/p>\n\n\n\n<div style=\"height:10px\" aria-hidden=\"true\" class=\"wp-block-spacer eplus-wrapper\"><\/div>\n\n\n\n<p class=\"eplus-wrapper wp-block-paragraph\">Using this framework, the authors compared explanation methods across several molecular machine learning architectures. Distance-aware transformer models, which explicitly encode interatomic geometry, consistently achieved higher PharmacoScore values than models that do not explicitly represent molecular geometry.<\/p>\n\n\n\n<div style=\"height:10px\" aria-hidden=\"true\" class=\"wp-block-spacer eplus-wrapper\"><\/div>\n\n\n\n<p class=\"eplus-wrapper wp-block-paragraph\"><em>High predictive accuracy does not necessarily mean that a model relies on chemically meaningful features. With PharmacoScore, we created a controlled way to test whether atom-level explanations recover the spatial relationships defined by pharmacophores. This gives us a common reference point for comparing explainability methods across different molecular machine learning architectures.<\/em><\/p>\n\n\n\n<p class=\"has-text-align-right eplus-wrapper wp-block-paragraph\">\u2014 Adam Su\u0142ek<\/p>\n\n\n\n<div style=\"height:10px\" aria-hidden=\"true\" class=\"wp-block-spacer eplus-wrapper\"><\/div>\n\n\n\n<p class=\"eplus-wrapper wp-block-paragraph\">The study provides a step towards molecular machine learning models that are evaluated not only by their predictive performance, but also by the chemical relevance of their explanations.<\/p>\n\n\n\n<div style=\"height:10px\" aria-hidden=\"true\" class=\"wp-block-spacer eplus-wrapper\"><\/div>\n\n\n\n<h3 id=\"h-congratulations-to-all-the-authors\" class=\"wp-block-heading eplus-wrapper\">Congratulations to all the authors!<\/h3>\n\n\n\n<div style=\"height:30px\" aria-hidden=\"true\" class=\"wp-block-spacer eplus-wrapper\"><\/div>\n\n\n\n<figure class=\"wp-block-image size-large eplus-wrapper\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"88\" src=\"https:\/\/sano.science\/wp-content\/uploads\/2026\/07\/FIRST-TEAM-FENG-FNP-Foundation-for-Polish-Science-1024x88.png\" alt=\"\" class=\"wp-image-32193\" srcset=\"https:\/\/sano.science\/wp-content\/uploads\/2026\/07\/FIRST-TEAM-FENG-FNP-Foundation-for-Polish-Science-1024x88.png 1024w, https:\/\/sano.science\/wp-content\/uploads\/2026\/07\/FIRST-TEAM-FENG-FNP-Foundation-for-Polish-Science-300x26.png 300w, https:\/\/sano.science\/wp-content\/uploads\/2026\/07\/FIRST-TEAM-FENG-FNP-Foundation-for-Polish-Science-768x66.png 768w, https:\/\/sano.science\/wp-content\/uploads\/2026\/07\/FIRST-TEAM-FENG-FNP-Foundation-for-Polish-Science-1536x132.png 1536w, https:\/\/sano.science\/wp-content\/uploads\/2026\/07\/FIRST-TEAM-FENG-FNP-Foundation-for-Polish-Science.png 1654w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><\/figure>\n\n\n\n<div style=\"height:30px\" aria-hidden=\"true\" class=\"wp-block-spacer eplus-wrapper\"><\/div>\n\n\n\n<p class=\"eplus-wrapper wp-block-paragraph\">This publication is one of the outcomes of the FIRST TEAM FENG project\u00a0<a href=\"https:\/\/www.linkedin.com\/company\/foundation-for-polish-science\/\" target=\"_blank\" rel=\"noreferrer noopener\">FNP Foundation for Polish Science<\/a>, under which\u00a0<a href=\"https:\/\/www.linkedin.com\/company\/lukasiewicz-krakow-institute-of-technology\/\" target=\"_blank\" rel=\"noreferrer noopener\">\u0141ukasiewicz \u2013 Krakow Institute of Technology<\/a>\u00a0is developing\u00a0new approaches\u00a0to the identification and evaluation of therapeutic candidates in hormone-dependent breast cancer.\u00a0<\/p>\n","protected":false},"excerpt":"New Publication in the Journal of Computational Science ","author":8,"featured_media":30902,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"inline_featured_image":false,"editor_plus_post_options":"{}","editor_plus_copied_stylings":"{}","footnotes":""},"categories":[1],"tags":[],"class_list":["post-32205","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-uncategorized"],"yoast_head":"<!-- This site is optimized with the Yoast SEO Premium plugin v28.1 (Yoast SEO v28.1) - https:\/\/yoast.com\/product\/yoast-seo-premium-wordpress\/ -->\n<title>Bringing clarity to how machine learning models &quot;see&quot; molecules\u00a0 - Centre for Computational Personalized Medicine<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/sano.science\/bringing-clarity-to-how-machine-learning-models-see-molecules\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Bringing clarity to how machine learning models &quot;see&quot; molecules\u00a0\" \/>\n<meta property=\"og:description\" content=\"New Publication in the Journal of Computational Science\" \/>\n<meta property=\"og:url\" content=\"https:\/\/sano.science\/bringing-clarity-to-how-machine-learning-models-see-molecules\/\" \/>\n<meta property=\"og:site_name\" content=\"Centre for Computational Personalized Medicine\" \/>\n<meta property=\"article:publisher\" content=\"https:\/\/www.facebook.com\/sano.science\/\" \/>\n<meta property=\"article:published_time\" content=\"2026-07-22T09:19:25+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2026-07-22T09:36:28+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/sano.science\/wp-content\/uploads\/2026\/04\/publications-Sano.jpg\" \/>\n\t<meta property=\"og:image:width\" content=\"1217\" \/>\n\t<meta property=\"og:image:height\" content=\"647\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/jpeg\" \/>\n<meta name=\"author\" content=\"Sano\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:creator\" content=\"@sanoscience\" \/>\n<meta name=\"twitter:site\" content=\"@sanoscience\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"Sano\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"2 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\\\/\\\/sano.science\\\/bringing-clarity-to-how-machine-learning-models-see-molecules\\\/#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/sano.science\\\/bringing-clarity-to-how-machine-learning-models-see-molecules\\\/\"},\"author\":{\"name\":\"Sano\",\"@id\":\"https:\\\/\\\/sano.science\\\/#\\\/schema\\\/person\\\/561b69b48b6a8f7904aed06df5d03c98\"},\"headline\":\"Bringing clarity to how machine learning models &#8220;see&#8221; molecules\u00a0\",\"datePublished\":\"2026-07-22T09:19:25+00:00\",\"dateModified\":\"2026-07-22T09:36:28+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/sano.science\\\/bringing-clarity-to-how-machine-learning-models-see-molecules\\\/\"},\"wordCount\":307,\"publisher\":{\"@id\":\"https:\\\/\\\/sano.science\\\/#organization\"},\"image\":{\"@id\":\"https:\\\/\\\/sano.science\\\/bringing-clarity-to-how-machine-learning-models-see-molecules\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/sano.science\\\/wp-content\\\/uploads\\\/2026\\\/04\\\/publications-Sano.jpg\",\"inLanguage\":\"en-US\"},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/sano.science\\\/bringing-clarity-to-how-machine-learning-models-see-molecules\\\/\",\"url\":\"https:\\\/\\\/sano.science\\\/bringing-clarity-to-how-machine-learning-models-see-molecules\\\/\",\"name\":\"Bringing clarity to how machine learning models \\\"see\\\" molecules\u00a0 - Centre for Computational Personalized Medicine\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/sano.science\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/sano.science\\\/bringing-clarity-to-how-machine-learning-models-see-molecules\\\/#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/sano.science\\\/bringing-clarity-to-how-machine-learning-models-see-molecules\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/sano.science\\\/wp-content\\\/uploads\\\/2026\\\/04\\\/publications-Sano.jpg\",\"datePublished\":\"2026-07-22T09:19:25+00:00\",\"dateModified\":\"2026-07-22T09:36:28+00:00\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/sano.science\\\/bringing-clarity-to-how-machine-learning-models-see-molecules\\\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/sano.science\\\/bringing-clarity-to-how-machine-learning-models-see-molecules\\\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/sano.science\\\/bringing-clarity-to-how-machine-learning-models-see-molecules\\\/#primaryimage\",\"url\":\"https:\\\/\\\/sano.science\\\/wp-content\\\/uploads\\\/2026\\\/04\\\/publications-Sano.jpg\",\"contentUrl\":\"https:\\\/\\\/sano.science\\\/wp-content\\\/uploads\\\/2026\\\/04\\\/publications-Sano.jpg\",\"width\":1217,\"height\":647,\"caption\":\"\\\"Sano Centre for Computational Medicine graphic used for news articles on scientific publications.\\\"\"},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/sano.science\\\/bringing-clarity-to-how-machine-learning-models-see-molecules\\\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/sano.science\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Bringing clarity to how machine learning models &#8220;see&#8221; molecules\u00a0\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/sano.science\\\/#website\",\"url\":\"https:\\\/\\\/sano.science\\\/\",\"name\":\"Centre for Computational Personalized Medicine\",\"description\":\"Sano \u2013 Centre for Computational Medicine\",\"publisher\":{\"@id\":\"https:\\\/\\\/sano.science\\\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/sano.science\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Organization\",\"@id\":\"https:\\\/\\\/sano.science\\\/#organization\",\"name\":\"Sano \u2013 Centre for Computational Medicine\",\"alternateName\":\"Sano\",\"url\":\"https:\\\/\\\/sano.science\\\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/sano.science\\\/#\\\/schema\\\/logo\\\/image\\\/\",\"url\":\"https:\\\/\\\/sano.science\\\/wp-content\\\/uploads\\\/2024\\\/05\\\/logo_sano_podstawowe.png\",\"contentUrl\":\"https:\\\/\\\/sano.science\\\/wp-content\\\/uploads\\\/2024\\\/05\\\/logo_sano_podstawowe.png\",\"width\":700,\"height\":265,\"caption\":\"Sano \u2013 Centre for Computational Medicine\"},\"image\":{\"@id\":\"https:\\\/\\\/sano.science\\\/#\\\/schema\\\/logo\\\/image\\\/\"},\"sameAs\":[\"https:\\\/\\\/www.facebook.com\\\/sano.science\\\/\",\"https:\\\/\\\/x.com\\\/sanoscience\",\"https:\\\/\\\/www.linkedin.com\\\/company\\\/sanoscience\\\/\",\"https:\\\/\\\/www.youtube.com\\\/channel\\\/UCDZ_8TcjMWUG2ZcgKKgfpwQ\",\"https:\\\/\\\/bsky.app\\\/profile\\\/sanoscience.bsky.social\"]},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/sano.science\\\/#\\\/schema\\\/person\\\/561b69b48b6a8f7904aed06df5d03c98\",\"name\":\"Sano\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/3349c3947cb21b81e9745a6442fe4ec43961bb077198eb416cbdc66ec45ea97c?s=96&d=mm&r=g\",\"url\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/3349c3947cb21b81e9745a6442fe4ec43961bb077198eb416cbdc66ec45ea97c?s=96&d=mm&r=g\",\"contentUrl\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/3349c3947cb21b81e9745a6442fe4ec43961bb077198eb416cbdc66ec45ea97c?s=96&d=mm&r=g\",\"caption\":\"Sano\"},\"url\":\"https:\\\/\\\/sano.science\\\/author\\\/s15to\\\/\"}]}<\/script>\n<!-- \/ Yoast SEO Premium plugin. -->","yoast_head_json":{"title":"Bringing clarity to how machine learning models \"see\" molecules\u00a0 - Centre for Computational Personalized Medicine","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/sano.science\/bringing-clarity-to-how-machine-learning-models-see-molecules\/","og_locale":"en_US","og_type":"article","og_title":"Bringing clarity to how machine learning models \"see\" molecules\u00a0","og_description":"New Publication in the Journal of Computational Science","og_url":"https:\/\/sano.science\/bringing-clarity-to-how-machine-learning-models-see-molecules\/","og_site_name":"Centre for Computational Personalized Medicine","article_publisher":"https:\/\/www.facebook.com\/sano.science\/","article_published_time":"2026-07-22T09:19:25+00:00","article_modified_time":"2026-07-22T09:36:28+00:00","og_image":[{"width":1217,"height":647,"url":"https:\/\/sano.science\/wp-content\/uploads\/2026\/04\/publications-Sano.jpg","type":"image\/jpeg"}],"author":"Sano","twitter_card":"summary_large_image","twitter_creator":"@sanoscience","twitter_site":"@sanoscience","twitter_misc":{"Written by":"Sano","Est. reading time":"2 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/sano.science\/bringing-clarity-to-how-machine-learning-models-see-molecules\/#article","isPartOf":{"@id":"https:\/\/sano.science\/bringing-clarity-to-how-machine-learning-models-see-molecules\/"},"author":{"name":"Sano","@id":"https:\/\/sano.science\/#\/schema\/person\/561b69b48b6a8f7904aed06df5d03c98"},"headline":"Bringing clarity to how machine learning models &#8220;see&#8221; molecules\u00a0","datePublished":"2026-07-22T09:19:25+00:00","dateModified":"2026-07-22T09:36:28+00:00","mainEntityOfPage":{"@id":"https:\/\/sano.science\/bringing-clarity-to-how-machine-learning-models-see-molecules\/"},"wordCount":307,"publisher":{"@id":"https:\/\/sano.science\/#organization"},"image":{"@id":"https:\/\/sano.science\/bringing-clarity-to-how-machine-learning-models-see-molecules\/#primaryimage"},"thumbnailUrl":"https:\/\/sano.science\/wp-content\/uploads\/2026\/04\/publications-Sano.jpg","inLanguage":"en-US"},{"@type":"WebPage","@id":"https:\/\/sano.science\/bringing-clarity-to-how-machine-learning-models-see-molecules\/","url":"https:\/\/sano.science\/bringing-clarity-to-how-machine-learning-models-see-molecules\/","name":"Bringing clarity to how machine learning models \"see\" molecules\u00a0 - Centre for Computational Personalized Medicine","isPartOf":{"@id":"https:\/\/sano.science\/#website"},"primaryImageOfPage":{"@id":"https:\/\/sano.science\/bringing-clarity-to-how-machine-learning-models-see-molecules\/#primaryimage"},"image":{"@id":"https:\/\/sano.science\/bringing-clarity-to-how-machine-learning-models-see-molecules\/#primaryimage"},"thumbnailUrl":"https:\/\/sano.science\/wp-content\/uploads\/2026\/04\/publications-Sano.jpg","datePublished":"2026-07-22T09:19:25+00:00","dateModified":"2026-07-22T09:36:28+00:00","breadcrumb":{"@id":"https:\/\/sano.science\/bringing-clarity-to-how-machine-learning-models-see-molecules\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/sano.science\/bringing-clarity-to-how-machine-learning-models-see-molecules\/"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/sano.science\/bringing-clarity-to-how-machine-learning-models-see-molecules\/#primaryimage","url":"https:\/\/sano.science\/wp-content\/uploads\/2026\/04\/publications-Sano.jpg","contentUrl":"https:\/\/sano.science\/wp-content\/uploads\/2026\/04\/publications-Sano.jpg","width":1217,"height":647,"caption":"\"Sano Centre for Computational Medicine graphic used for news articles on scientific publications.\""},{"@type":"BreadcrumbList","@id":"https:\/\/sano.science\/bringing-clarity-to-how-machine-learning-models-see-molecules\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/sano.science\/"},{"@type":"ListItem","position":2,"name":"Bringing clarity to how machine learning models &#8220;see&#8221; molecules\u00a0"}]},{"@type":"WebSite","@id":"https:\/\/sano.science\/#website","url":"https:\/\/sano.science\/","name":"Centre for Computational Personalized Medicine","description":"Sano \u2013 Centre for Computational Medicine","publisher":{"@id":"https:\/\/sano.science\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/sano.science\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Organization","@id":"https:\/\/sano.science\/#organization","name":"Sano \u2013 Centre for Computational Medicine","alternateName":"Sano","url":"https:\/\/sano.science\/","logo":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/sano.science\/#\/schema\/logo\/image\/","url":"https:\/\/sano.science\/wp-content\/uploads\/2024\/05\/logo_sano_podstawowe.png","contentUrl":"https:\/\/sano.science\/wp-content\/uploads\/2024\/05\/logo_sano_podstawowe.png","width":700,"height":265,"caption":"Sano \u2013 Centre for Computational Medicine"},"image":{"@id":"https:\/\/sano.science\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/www.facebook.com\/sano.science\/","https:\/\/x.com\/sanoscience","https:\/\/www.linkedin.com\/company\/sanoscience\/","https:\/\/www.youtube.com\/channel\/UCDZ_8TcjMWUG2ZcgKKgfpwQ","https:\/\/bsky.app\/profile\/sanoscience.bsky.social"]},{"@type":"Person","@id":"https:\/\/sano.science\/#\/schema\/person\/561b69b48b6a8f7904aed06df5d03c98","name":"Sano","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/secure.gravatar.com\/avatar\/3349c3947cb21b81e9745a6442fe4ec43961bb077198eb416cbdc66ec45ea97c?s=96&d=mm&r=g","url":"https:\/\/secure.gravatar.com\/avatar\/3349c3947cb21b81e9745a6442fe4ec43961bb077198eb416cbdc66ec45ea97c?s=96&d=mm&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/3349c3947cb21b81e9745a6442fe4ec43961bb077198eb416cbdc66ec45ea97c?s=96&d=mm&r=g","caption":"Sano"},"url":"https:\/\/sano.science\/author\/s15to\/"}]}},"acf":[],"gutenberg_blocks":[{"blockName":"custom-styles","attrs":{"styles":""}},{"blockName":"core\/paragraph","attrs":{"epAnimationGeneratedClass":"edplus_anim-ht9Ll0","epGeneratedClass":"eplus-wrapper"},"innerBlocks":[],"innerHTML":"\n<p class=\" eplus-wrapper\">Our researchers Adam Su\u0142ek and Tomasz Ko\u015bcio\u0142ek, together with their co-authors, have published a new paper in the <em>Journal of Computational Science<\/em>, addressing one of the key challenges in molecular machine learning: evaluating atom-level explanations.<\/p>\n","innerContent":["\n<p class=\" eplus-wrapper\">Our researchers Adam Su\u0142ek and Tomasz Ko\u015bcio\u0142ek, together with their co-authors, have published a new paper in the <em>Journal of Computational Science<\/em>, addressing one of the key challenges in molecular machine learning: evaluating atom-level explanations.<\/p>\n"]},{"blockName":"core\/spacer","attrs":{"height":"10px","epAnimationGeneratedClass":"edplus_anim-FzIAQF","epGeneratedClass":"eplus-wrapper"},"innerBlocks":[],"innerHTML":"\n<div style=\"height:10px\" aria-hidden=\"true\" class=\"wp-block-spacer eplus-wrapper\"><\/div>\n","innerContent":["\n<div style=\"height:10px\" aria-hidden=\"true\" class=\"wp-block-spacer eplus-wrapper\"><\/div>\n"]},{"blockName":"core\/heading","attrs":{"anchor":"h-benchmarking-atom-level-explainability-against-pharmacophore-computed-labels-in-molecular-machine-learning","epAnimationGeneratedClass":"edplus_anim-1Oeht0","epGeneratedClass":"eplus-wrapper"},"innerBlocks":[],"innerHTML":"\n<h2 id=\"h-benchmarking-atom-level-explainability-against-pharmacophore-computed-labels-in-molecular-machine-learning\" class=\"wp-block-heading eplus-wrapper\">\u201cBenchmarking atom-level explainability against pharmacophore-computed labels in molecular machine learning\u201d<\/h2>\n","innerContent":["\n<h2 id=\"h-benchmarking-atom-level-explainability-against-pharmacophore-computed-labels-in-molecular-machine-learning\" class=\"wp-block-heading eplus-wrapper\">\u201cBenchmarking atom-level explainability against pharmacophore-computed labels in molecular machine learning\u201d<\/h2>\n"]},{"blockName":"core\/paragraph","attrs":{"epAnimationGeneratedClass":"edplus_anim-xO2Vh2","epGeneratedClass":"eplus-wrapper"},"innerBlocks":[],"innerHTML":"\n<p class=\" eplus-wrapper\"><strong>Authors:<\/strong> Adam Su\u0142ek, Jakub Klimczak, Jakub Jo\u0144czyk, Tomasz Ko\u015bcio\u0142ek, Tomasz Danel and Barbara Pucelik<\/p>\n","innerContent":["\n<p class=\" eplus-wrapper\"><strong>Authors:<\/strong> Adam Su\u0142ek, Jakub Klimczak, Jakub Jo\u0144czyk, Tomasz Ko\u015bcio\u0142ek, Tomasz Danel and Barbara Pucelik<\/p>\n"]},{"blockName":"core\/spacer","attrs":{"height":"10px","epAnimationGeneratedClass":"edplus_anim-SdkLAi","epGeneratedClass":"eplus-wrapper"},"innerBlocks":[],"innerHTML":"\n<div style=\"height:10px\" aria-hidden=\"true\" class=\"wp-block-spacer eplus-wrapper\"><\/div>\n","innerContent":["\n<div style=\"height:10px\" aria-hidden=\"true\" class=\"wp-block-spacer eplus-wrapper\"><\/div>\n"]},{"blockName":"core\/paragraph","attrs":{"epAnimationGeneratedClass":"edplus_anim-qqP5xk","epGeneratedClass":"eplus-wrapper"},"innerBlocks":[],"innerHTML":"\n<p class=\" eplus-wrapper\">Machine learning models can predict molecular properties with impressive accuracy\u2014but do their explanations identify atoms that are genuinely relevant to molecular structure and function?<\/p>\n","innerContent":["\n<p class=\" eplus-wrapper\">Machine learning models can predict molecular properties with impressive accuracy\u2014but do their explanations identify atoms that are genuinely relevant to molecular structure and function?<\/p>\n"]},{"blockName":"core\/spacer","attrs":{"height":"10px","epAnimationGeneratedClass":"edplus_anim-SdkLAi","epGeneratedClass":"eplus-wrapper"},"innerBlocks":[],"innerHTML":"\n<div style=\"height:10px\" aria-hidden=\"true\" class=\"wp-block-spacer eplus-wrapper\"><\/div>\n","innerContent":["\n<div style=\"height:10px\" aria-hidden=\"true\" class=\"wp-block-spacer eplus-wrapper\"><\/div>\n"]},{"blockName":"core\/paragraph","attrs":{"epAnimationGeneratedClass":"edplus_anim-3eMRfO","epGeneratedClass":"eplus-wrapper"},"innerBlocks":[],"innerHTML":"\n<p class=\" eplus-wrapper\">The study introduces a controlled benchmark based on pharmacophores and geometric rules describing the spatial arrangement of functional elements within molecules. At the core of the framework is PharmacoScore, a metric that quantifies how closely atom-level model explanations align with pharmacophore-derived reference annotations.<\/p>\n","innerContent":["\n<p class=\" eplus-wrapper\">The study introduces a controlled benchmark based on pharmacophores and geometric rules describing the spatial arrangement of functional elements within molecules. At the core of the framework is PharmacoScore, a metric that quantifies how closely atom-level model explanations align with pharmacophore-derived reference annotations.<\/p>\n"]},{"blockName":"core\/spacer","attrs":{"height":"10px","epAnimationGeneratedClass":"edplus_anim-SdkLAi","epGeneratedClass":"eplus-wrapper"},"innerBlocks":[],"innerHTML":"\n<div style=\"height:10px\" aria-hidden=\"true\" class=\"wp-block-spacer eplus-wrapper\"><\/div>\n","innerContent":["\n<div style=\"height:10px\" aria-hidden=\"true\" class=\"wp-block-spacer eplus-wrapper\"><\/div>\n"]},{"blockName":"core\/paragraph","attrs":{"epAnimationGeneratedClass":"edplus_anim-1wBFSj","epGeneratedClass":"eplus-wrapper"},"innerBlocks":[],"innerHTML":"\n<p class=\" eplus-wrapper\">Using this framework, the authors compared explanation methods across several molecular machine learning architectures. Distance-aware transformer models, which explicitly encode interatomic geometry, consistently achieved higher PharmacoScore values than models that do not explicitly represent molecular geometry.<\/p>\n","innerContent":["\n<p class=\" eplus-wrapper\">Using this framework, the authors compared explanation methods across several molecular machine learning architectures. Distance-aware transformer models, which explicitly encode interatomic geometry, consistently achieved higher PharmacoScore values than models that do not explicitly represent molecular geometry.<\/p>\n"]},{"blockName":"core\/spacer","attrs":{"height":"10px","epAnimationGeneratedClass":"edplus_anim-SdkLAi","epGeneratedClass":"eplus-wrapper"},"innerBlocks":[],"innerHTML":"\n<div style=\"height:10px\" aria-hidden=\"true\" class=\"wp-block-spacer eplus-wrapper\"><\/div>\n","innerContent":["\n<div style=\"height:10px\" aria-hidden=\"true\" class=\"wp-block-spacer eplus-wrapper\"><\/div>\n"]},{"blockName":"core\/paragraph","attrs":{"epAnimationGeneratedClass":"edplus_anim-cyJSbF","epGeneratedClass":"eplus-wrapper"},"innerBlocks":[],"innerHTML":"\n<p class=\" eplus-wrapper\"><em>High predictive accuracy does not necessarily mean that a model relies on chemically meaningful features. With PharmacoScore, we created a controlled way to test whether atom-level explanations recover the spatial relationships defined by pharmacophores. This gives us a common reference point for comparing explainability methods across different molecular machine learning architectures.<\/em><\/p>\n","innerContent":["\n<p class=\" eplus-wrapper\"><em>High predictive accuracy does not necessarily mean that a model relies on chemically meaningful features. With PharmacoScore, we created a controlled way to test whether atom-level explanations recover the spatial relationships defined by pharmacophores. This gives us a common reference point for comparing explainability methods across different molecular machine learning architectures.<\/em><\/p>\n"]},{"blockName":"core\/paragraph","attrs":{"style":{"typography":{"textAlign":"right"}},"epAnimationGeneratedClass":"edplus_anim-CWBNpJ","epGeneratedClass":"eplus-wrapper"},"innerBlocks":[],"innerHTML":"\n<p class=\"has-text-align-right eplus-wrapper\">\u2014 Adam Su\u0142ek<\/p>\n","innerContent":["\n<p class=\"has-text-align-right eplus-wrapper\">\u2014 Adam Su\u0142ek<\/p>\n"]},{"blockName":"core\/spacer","attrs":{"height":"10px","epAnimationGeneratedClass":"edplus_anim-SdkLAi","epGeneratedClass":"eplus-wrapper"},"innerBlocks":[],"innerHTML":"\n<div style=\"height:10px\" aria-hidden=\"true\" class=\"wp-block-spacer eplus-wrapper\"><\/div>\n","innerContent":["\n<div style=\"height:10px\" aria-hidden=\"true\" class=\"wp-block-spacer eplus-wrapper\"><\/div>\n"]},{"blockName":"core\/paragraph","attrs":{"epAnimationGeneratedClass":"edplus_anim-yYBUmh","epGeneratedClass":"eplus-wrapper"},"innerBlocks":[],"innerHTML":"\n<p class=\" eplus-wrapper\">The study provides a step towards molecular machine learning models that are evaluated not only by their predictive performance, but also by the chemical relevance of their explanations.<\/p>\n","innerContent":["\n<p class=\" eplus-wrapper\">The study provides a step towards molecular machine learning models that are evaluated not only by their predictive performance, but also by the chemical relevance of their explanations.<\/p>\n"]},{"blockName":"core\/spacer","attrs":{"height":"10px","epAnimationGeneratedClass":"edplus_anim-SdkLAi","epGeneratedClass":"eplus-wrapper"},"innerBlocks":[],"innerHTML":"\n<div style=\"height:10px\" aria-hidden=\"true\" class=\"wp-block-spacer eplus-wrapper\"><\/div>\n","innerContent":["\n<div style=\"height:10px\" aria-hidden=\"true\" class=\"wp-block-spacer eplus-wrapper\"><\/div>\n"]},{"blockName":"core\/heading","attrs":{"level":3,"anchor":"h-congratulations-to-all-the-authors","epAnimationGeneratedClass":"edplus_anim-vKY4ON","epGeneratedClass":"eplus-wrapper"},"innerBlocks":[],"innerHTML":"\n<h3 id=\"h-congratulations-to-all-the-authors\" class=\"wp-block-heading eplus-wrapper\">Congratulations to all the authors!<\/h3>\n","innerContent":["\n<h3 id=\"h-congratulations-to-all-the-authors\" class=\"wp-block-heading eplus-wrapper\">Congratulations to all the authors!<\/h3>\n"]},{"blockName":"core\/spacer","attrs":{"height":"30px","epAnimationGeneratedClass":"edplus_anim-cuLNnq","epGeneratedClass":"eplus-wrapper"},"innerBlocks":[],"innerHTML":"\n<div style=\"height:30px\" aria-hidden=\"true\" class=\"wp-block-spacer eplus-wrapper\"><\/div>\n","innerContent":["\n<div style=\"height:30px\" aria-hidden=\"true\" class=\"wp-block-spacer eplus-wrapper\"><\/div>\n"]},{"blockName":"core\/image","attrs":{"id":32193,"sizeSlug":"large","linkDestination":"none","epAnimationGeneratedClass":"edplus_anim-d3YteN","epGeneratedClass":"eplus-wrapper"},"innerBlocks":[],"innerHTML":"\n<figure class=\"wp-block-image size-large eplus-wrapper\"><img src=\"https:\/\/sano.science\/wp-content\/uploads\/2026\/07\/FIRST-TEAM-FENG-FNP-Foundation-for-Polish-Science-1024x88.png\" alt=\"\" class=\"wp-image-32193\"\/><\/figure>\n","innerContent":["\n<figure class=\"wp-block-image size-large eplus-wrapper\"><img src=\"https:\/\/sano.science\/wp-content\/uploads\/2026\/07\/FIRST-TEAM-FENG-FNP-Foundation-for-Polish-Science-1024x88.png\" alt=\"\" class=\"wp-image-32193\"\/><\/figure>\n"]},{"blockName":"core\/spacer","attrs":{"height":"30px","epAnimationGeneratedClass":"edplus_anim-qIJWPG","epGeneratedClass":"eplus-wrapper"},"innerBlocks":[],"innerHTML":"\n<div style=\"height:30px\" aria-hidden=\"true\" class=\"wp-block-spacer eplus-wrapper\"><\/div>\n","innerContent":["\n<div style=\"height:30px\" aria-hidden=\"true\" class=\"wp-block-spacer eplus-wrapper\"><\/div>\n"]},{"blockName":"core\/paragraph","attrs":{"epAnimationGeneratedClass":"edplus_anim-r14Vbx","epGeneratedClass":"eplus-wrapper"},"innerBlocks":[],"innerHTML":"\n<p class=\" eplus-wrapper\">This publication is one of the outcomes of the FIRST TEAM FENG project\u00a0<a href=\"https:\/\/www.linkedin.com\/company\/foundation-for-polish-science\/\" target=\"_blank\" rel=\"noreferrer noopener\">FNP Foundation for Polish Science<\/a>, under which\u00a0<a href=\"https:\/\/www.linkedin.com\/company\/lukasiewicz-krakow-institute-of-technology\/\" target=\"_blank\" rel=\"noreferrer noopener\">\u0141ukasiewicz \u2013 Krakow Institute of Technology<\/a>\u00a0is developing\u00a0new approaches\u00a0to the identification and evaluation of therapeutic candidates in hormone-dependent breast cancer.\u00a0<\/p>\n","innerContent":["\n<p class=\" eplus-wrapper\">This publication is one of the outcomes of the FIRST TEAM FENG project\u00a0<a href=\"https:\/\/www.linkedin.com\/company\/foundation-for-polish-science\/\" target=\"_blank\" rel=\"noreferrer noopener\">FNP Foundation for Polish Science<\/a>, under which\u00a0<a href=\"https:\/\/www.linkedin.com\/company\/lukasiewicz-krakow-institute-of-technology\/\" target=\"_blank\" rel=\"noreferrer noopener\">\u0141ukasiewicz \u2013 Krakow Institute of Technology<\/a>\u00a0is developing\u00a0new approaches\u00a0to the identification and evaluation of therapeutic candidates in hormone-dependent breast cancer.\u00a0<\/p>\n"]}],"meta_data":{"has_thumbnail_pattern":false,"share_on_social_media":{"has_social_media":false}},"featured_image":{"url":"https:\/\/sano.science\/wp-content\/uploads\/2026\/04\/publications-Sano-1024x544.jpg"},"main_category":{"name":"Uncategorized"},"prev_page":{"slug":"new-issue-of-sano-news-is-out-2"},"next_page":{"slug":"sano-at-bit26-tracking-the-sinus-microbiome"},"_links":{"self":[{"href":"https:\/\/sano.science\/index.php\/wp-json\/wp\/v2\/posts\/32205","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/sano.science\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/sano.science\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/sano.science\/index.php\/wp-json\/wp\/v2\/users\/8"}],"replies":[{"embeddable":true,"href":"https:\/\/sano.science\/index.php\/wp-json\/wp\/v2\/comments?post=32205"}],"version-history":[{"count":5,"href":"https:\/\/sano.science\/index.php\/wp-json\/wp\/v2\/posts\/32205\/revisions"}],"predecessor-version":[{"id":32229,"href":"https:\/\/sano.science\/index.php\/wp-json\/wp\/v2\/posts\/32205\/revisions\/32229"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/sano.science\/index.php\/wp-json\/wp\/v2\/media\/30902"}],"wp:attachment":[{"href":"https:\/\/sano.science\/index.php\/wp-json\/wp\/v2\/media?parent=32205"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/sano.science\/index.php\/wp-json\/wp\/v2\/categories?post=32205"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/sano.science\/index.php\/wp-json\/wp\/v2\/tags?post=32205"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}