{"689835":{"#nid":"689835","#data":{"type":"news","title":"AI is Reengineering Drug Discovery by Speeding Up Testing and Scanning Petabytes of Data for Connections Between\u00a0Diseases","body":[{"value":"\u003Cdiv class=\u0022theconversation-article-body\u0022\u003E\u003Cp\u003E\u003Cem\u003EIn December, The Conversation hosted a webinar on AI\u2019s revolutionary role in drug discovery and development.\u003C\/em\u003E\u003C\/p\u003E\u003Cp\u003E\u003Cem\u003EScience and technology editor \u003C\/em\u003E\u003Ca href=\u0022https:\/\/theconversation.com\/profiles\/eric-smalley-944964\u0022\u003E\u003Cem\u003EEric Smalley\u003C\/em\u003E\u003C\/a\u003E\u003Cem\u003E interviewed \u003C\/em\u003E\u003Ca href=\u0022https:\/\/biosciences.gatech.edu\/people\/jeffrey-skolnick\u0022\u003E\u003Cem\u003EJeffrey Skolnick\u003C\/em\u003E\u003C\/a\u003E\u003Cem\u003E, eminent scholar in computational systems biology at Georgia Institute of Technology, and \u003C\/em\u003E\u003Ca href=\u0022https:\/\/medschool.vanderbilt.edu\/pharmacology\/person\/ben-brown\/\u0022\u003E\u003Cem\u003EBenjamin P. Brown\u003C\/em\u003E\u003C\/a\u003E\u003Cem\u003E, assistant professor of pharmacology at Vanderbilt University.\u003C\/em\u003E\u003C\/p\u003E\u003Cp\u003E\u003Cem\u003ESkolnick has developed AI-based approaches to predict protein structure and function that may help with drug discovery and finding off-label uses of existing drugs. Brown\u2019s lab works on creating new computer models that make drug discovery faster and more reliable. Below is a condensed and edited version of the interview.\u003C\/em\u003E\u003C\/p\u003E\u003Ch4\u003E\u003Cstrong\u003ELet\u2019s start with the big picture. How is AI changing biomedical research and drug discovery, and what is the potential we are talking about?\u003C\/strong\u003E\u003C\/h4\u003E\u003Cp\u003E\u003Cstrong\u003ESkolnick:\u003C\/strong\u003E The upside, potentially, is very large. One of the frustrating things about drug discovery is that, in spite of the fact that the people doing it are extraordinarily intelligent and have done an extraordinarily good job, \u003Ca href=\u0022https:\/\/doi.org\/10.1016\/j.apsb.2022.02.002\u0022\u003Ethe success rate is very low\u003C\/a\u003E. About \u003Ca href=\u0022https:\/\/doi.org\/10.1016\/j.apsb.2022.02.002\u0022\u003E1 in 5\u003C\/a\u003E drugs will have negative health effects that outweigh its benefits. Of the ones that pass, \u003Ca href=\u0022https:\/\/doi.org\/10.1016\/j.apsb.2022.02.002\u0022\u003Eroughly half don\u2019t work\u003C\/a\u003E.\u003C\/p\u003E\u003Cp\u003EIn drug development, there are several key issues: Can you predict which target is driving a particular disease? Once this target is identified, how can you guarantee the drug is going to work and isn\u2019t simultaneously going to kill you?\u003C\/p\u003E\u003Cp\u003EThese are outstanding problems in drug discovery in which AI can play an important, though not 100% guaranteed, role. Unlike us, AI can look at basically \u003Ca href=\u0022https:\/\/academic.oup.com\/nsr\/article\/12\/5\/nwaf050\/8029900\u0022\u003Eall available knowledge\u003C\/a\u003E. On a good day it makes strong and true connections called \u201c\u003Ca href=\u0022https:\/\/doi.org\/10.1016\/bs.adcom.2023.02.001\u0022\u003Einsights\u003C\/a\u003E,\u201d and on a bad day it does what is called \u201c\u003Ca href=\u0022https:\/\/theconversation.com\/what-are-ai-hallucinations-why-ais-sometimes-make-things-up-242896\u0022\u003Ehallucinating\u003C\/a\u003E\u201d and sees things that are weak and probably false.\u003C\/p\u003E\u003Cfigure\u003E\u003Cp\u003E\u003Ciframe width=\u0022440\u0022 height=\u0022260\u0022 src=\u0022https:\/\/www.youtube.com\/embed\/lHC_9x3IXZ0?wmode=transparent\u0026amp;start=0\u0022 frameborder=\u00220\u0022 allowfullscreen=\u0022\u0022\u003E\u003C\/iframe\u003E\u003C\/p\u003E\u003Cfigcaption\u003E\u003Cspan class=\u0022caption\u0022\u003EEric Smalley interviews Jeffrey Skolnick and Benjamin P. Brown.\u003C\/span\u003E\u003C\/figcaption\u003E\u003C\/figure\u003E\u003Cp\u003EAt the end of the day, many diseases do not have a cure. Most diseases are maintained, such as high cholesterol or autoimmune conditions. A treatment for cancer might buy you five years, and now you\u2019re in Stage 4 and you\u2019ve exhausted all the standard care drugs. \u003Ca href=\u0022https:\/\/doi.org\/10.3390\/ph16060891\u0022\u003EAI can play a role\u003C\/a\u003E to suggest alternatives where there are none.\u003C\/p\u003E\u003Ch4\u003E\u003Cstrong\u003ELet\u2019s give some basic definitions here. When we use the word drug, we\u2019re talking about a wide range of therapies. Can you explain the range \u2013 we\u2019ve got small molecule drugs, biologics, gene therapies, cell therapies.\u003C\/strong\u003E\u003C\/h4\u003E\u003Cp\u003E\u003Cstrong\u003EBrown:\u003C\/strong\u003E We have fairly large molecules in our bodies called proteins. They are like machines that \u003Ca href=\u0022https:\/\/www.ncbi.nlm.nih.gov\/books\/NBK26911\/\u0022\u003Ecarry out specific functions\u003C\/a\u003E and interact with one another. Oftentimes, when we\u2019re trying to treat disease, we\u2019re trying to \u003Ca href=\u0022https:\/\/doi.org\/10.1002\/mco2.261\u0022\u003Ealter functions of specific proteins\u003C\/a\u003E. Many drugs, like \u003Ca href=\u0022https:\/\/doi.org\/10.1016\/S0049-3848(03)00379-7\u0022\u003Easpirin\u003C\/a\u003E and \u003Ca href=\u0022https:\/\/doi.org\/10.1086\/317517\u0022\u003ETylenol\u003C\/a\u003E, are small molecules that can fit into a protein and change its function. Fundamentally, drugs don\u2019t have to just interact with proteins, but this is a major way in which our current repertoire of medications work.\u003C\/p\u003E\u003Cp\u003EThere are also proteins that act like drugs, such as \u003Ca href=\u0022https:\/\/doi.org\/10.1111\/imr.13387\u0022\u003Eantibodies\u003C\/a\u003E. When you receive a vaccine for a virus, your body is basically given \u003Ca href=\u0022https:\/\/doi.org\/10.1016\/B978-0-12-802174-3.00002-3\u0022\u003Einstructions on how to develop antibodies\u003C\/a\u003E. These antibodies will target some part of that virus. Your body is creating these big molecules, much bigger than aspirin, to go and interact with foreign proteins in a different way. \u003Ca href=\u0022https:\/\/doi.org\/10.1590\/S1679-45082017RB4024\u0022\u003EGene therapy\u003C\/a\u003E is a larger step beyond that.\u003C\/p\u003E\u003Cp\u003ESo these modalities \u2013 molecule, protein, antibody or gene \u2013 are very different types of molecules. They have different scales and rules, so the way you approach designing and discovering them various widely.\u003C\/p\u003E\u003Ch4\u003E\u003Cstrong\u003ECan you briefly explain artificial neural networks, and what the \u201cdeep\u201d in deep learning means?\u003C\/strong\u003E\u003C\/h4\u003E\u003Cp\u003E\u003Cstrong\u003ESkolnick:\u003C\/strong\u003E AlphaFold, developed by DeepMind, involved understanding how neural networks worked. They built a network with a lot of \u003Ca href=\u0022https:\/\/doi.org\/10.3390\/diagnostics13152582\u0022\u003Einputs, which are stimuli, and outputs with different weights\u003C\/a\u003E, similar to how your brain actually works. These simple connections, or neurons, have \u003Ca href=\u0022https:\/\/theconversation.com\/what-is-reinforcement-learning-an-ai-researcher-explains-a-key-method-of-teaching-machines-and-how-it-relates-to-training-your-dog-251887\u0022\u003Ereinforcement learning\u003C\/a\u003E.\u003C\/p\u003E\u003Cp\u003EThey also created sophisticated neural networks, such as \u003Ca href=\u0022https:\/\/doi.org\/10.1073\/pnas.2219150120\u0022\u003Etransformers, which do specific things\u003C\/a\u003E like a special-purpose tool that can learn, and they added a mechanism called \u201cattention,\u201d which \u003Ca href=\u0022https:\/\/doi.org\/10.1016\/j.inffus.2024.102417\u0022\u003Eamplifies critical details\u003C\/a\u003E. Super neural networks with transformers is what we call deep learning. These now have literally billions, if not trillions, of parameters.\u003C\/p\u003E\u003Cp\u003EEssentially, these machines \u003Ca href=\u0022https:\/\/doi.org\/10.52202\/079017-2495\u0022\u003Ecan learn higher order correlations between events\u003C\/a\u003E, meaning the patterns of conditional interactions that depend on the properties of multiple things simultaneously. In these higher order correlations, AI has the potential to see previously unknown things that are embedded in petabytes (a unit of data equivalent to \u003Ca href=\u0022https:\/\/www.eecis.udel.edu\/%7Eamer\/Table-Kilo-Mega-Giga---YottaBytes.html\u0022\u003Ehalf of the contents of all U.S. academic research libraries\u003C\/a\u003E of biological data.\u003C\/p\u003E\u003Cp\u003EAlphaFold, which \u003Ca href=\u0022https:\/\/doi.org\/10.1080\/14789450.2025.2456046\u0022\u003Epredicts three-dimensional, bioactive forms of a protein\u003C\/a\u003E, has millions of sequences and a couple of hundred thousand structures. It can tell you, based on a particular pattern, what \u003Ca href=\u0022https:\/\/doi.org\/10.3390\/ijms26146807\u0022\u003Esmall molecule to design\u003C\/a\u003E that sticks to a protein to induce some kind of structural shift.\u003C\/p\u003E\u003Ch4\u003E\u003Cstrong\u003EHow is this technology being used in biomedical research to understand molecular dynamics or, essentially, the biological processes involved in health and disease?\u003C\/strong\u003E\u003C\/h4\u003E\u003Cp\u003E\u003Cstrong\u003EBrown:\u003C\/strong\u003E In 2013, there was a Nobel Prize for \u003Ca href=\u0022https:\/\/doi.org\/10.1016\/j.str.2013.11.005\u0022\u003Emolecular dynamics simulations\u003C\/a\u003E, computational tools that help you understand the motions of molecules as they move according to physics. There\u2019s a huge body of scientific research built around those ideas.\u003C\/p\u003E\u003Cp\u003EAI and deep learning are large right now, but it\u2019s worth mentioning that for the last decade and a half, people have been \u003Ca href=\u0022https:\/\/doi.org\/10.1038\/nchembio.576\u0022\u003Eusing much smaller machine learning algorithms\u003C\/a\u003E to help design drugs. A lot of the ideas, such as [using machine learning for virtual screening], are not new and have been in practice for a while.\u003C\/p\u003E\u003Cp\u003EWith AlphaFold\u2019s technologies to help people design proteins and predict their structure, we\u2019ve changed how we think about a lot of these problems. We have this \u003Ca href=\u0022https:\/\/doi.org\/10.1016\/j.omtn.2024.102295\u0022\u003Enew repertoire of approaches\u003C\/a\u003E to build ideas around and to start thinking about drug discovery.\u003C\/p\u003E\u003Ch4\u003E\u003Cstrong\u003EFrom 20 years ago to now, what has today\u2019s AI technology done in terms of scale of change in this process?\u003C\/strong\u003E\u003C\/h4\u003E\u003Cp\u003E\u003Cstrong\u003ESkolnick:\u003C\/strong\u003E A lot of diseases, like cancers, are \u003Ca href=\u0022https:\/\/doi.org\/10.15430\/JCP.2018.23.4.153\u0022\u003Ecaused by a collection of malfunctioning proteins\u003C\/a\u003E. AI now allows us to start to think conceptually about how these diseases are organized and related to each other.\u003C\/p\u003E\u003Cp\u003EDiseases tend to co-occur. For example, if you have \u003Ca href=\u0022https:\/\/doi.org\/10.3389\/fendo.2024.1354372\u0022\u003Ehyperthyroidism, you\u2019re very likely to develop Alzheimer\u2019s\u003C\/a\u003E. Kind of weird, right? We can look at pieces, but AI can look at all the information, integrate the collective behavior and then identify common drivers. This allows you to construct disease interrelationships which offer the \u003Ca href=\u0022https:\/\/doi.org\/10.1002\/adtp.202300332\u0022\u003Epossibility of broad spectrum treatments\u003C\/a\u003E that \u003Ca href=\u0022https:\/\/www.nih.gov\/news-events\/nih-research-matters\/progress-toward-broad-spectrum-antiviral\u0022\u003Ecould treat whole collections of diseases\u003C\/a\u003E rather than narrow-spectrum treatments.\u003C\/p\u003E\u003Cp\u003ERelatedly, AI also can help us \u003Ca href=\u0022https:\/\/doi.org\/10.1002\/cpt.3153\u0022\u003Eunderstand disease trajectories\u003C\/a\u003E. Diseases that tend to \u003Ca href=\u0022https:\/\/doi.org\/10.1146\/annurev-biodatasci-110123-041001\u0022\u003Eco-occur often present themselves consecutively\u003C\/a\u003E. You have disease 1, it gives you disease 2, then gives you disease 3. This suggests that if you go back to the root with disease 1, you may be able to stop a whole bunch of stuff. You can\u2019t analyze millions of trajectories and millions of data without a tool, so you couldn\u2019t do this before.\u003C\/p\u003E\u003Cp\u003EThis holds a lot of promise, but one also must be careful not to overpromise. It will help, it will accelerate, but \u003Ca href=\u0022https:\/\/www.scienceopen.com\/hosted-document?doi=10.15212\/bioi-2025-0188\u0022\u003Eit is not a substitute yet for real experiments\u003C\/a\u003E, real clinical validation and trials.\u003C!-- Below is The Conversation\u0027s page counter tag. Please DO NOT REMOVE. --\u003E\u003Cimg style=\u0022border-color:!important;border-style:none;box-shadow:none !important;margin:0 !important;max-height:1px !important;max-width:1px !important;min-height:1px !important;min-width:1px !important;opacity:0 !important;outline:none !important;padding:0 !important;\u0022 src=\u0022https:\/\/counter.theconversation.com\/content\/274693\/count.gif?distributor=republish-lightbox-basic\u0022 alt=\u0022The Conversation\u0022 width=\u00221\u0022 height=\u00221\u0022 referrerpolicy=\u0022no-referrer-when-downgrade\u0022\u003E\u003C!-- End of code. If you don\u0027t see any code above, please get new code from the Advanced tab after you click the republish button. The page counter does not collect any personal data. More info: https:\/\/theconversation.com\/republishing-guidelines --\u003E\u003C\/p\u003E\u003Cp\u003E\u0026nbsp;\u003C\/p\u003E\u003Cp\u003E\u003Cem\u003EThis article is republished from \u003C\/em\u003E\u003Ca href=\u0022https:\/\/theconversation.com\u0022\u003E\u003Cem\u003EThe Conversation\u003C\/em\u003E\u003C\/a\u003E\u003Cem\u003E under a Creative Commons license. Read the \u003C\/em\u003E\u003Ca href=\u0022https:\/\/theconversation.com\/ai-is-reengineering-drug-discovery-by-speeding-up-testing-and-scanning-petabytes-of-data-for-connections-between-diseases-274693\u0022\u003E\u003Cem\u003Eoriginal article\u003C\/em\u003E\u003C\/a\u003E\u003Cem\u003E.\u003C\/em\u003E\u003C\/p\u003E\u003C\/div\u003E","summary":"","format":"full_html"}],"field_subtitle":"","field_summary":[{"value":"\u003Cp\u003EAI and machine learning provide new tools for scientists to think about drug discovery.\u003C\/p\u003E","format":"limited_html"}],"field_summary_sentence":[{"value":"AI and machine learning provide new tools for scientists to think about drug discovery."}],"uid":"27469","created_gmt":"2026-04-17 15:55:09","changed_gmt":"2026-04-17 16:14:02","author":"Kristen Bailey","boilerplate_text":"","field_publication":"","field_article_url":"","location":"Atlanta, GA","dateline":{"date":"2026-04-07T00:00:00-04:00","iso_date":"2026-04-07T00:00:00-04:00","tz":"America\/New_York"},"extras":[],"hg_media":{"679992":{"id":"679992","type":"image","title":" AI and machine learning provide new tools for scientists to think about drug discovery. gorodenkoff\/iStock via Getty Images ","body":"\u003Cp\u003E\u0026nbsp;AI and machine learning provide new tools for scientists to think about drug discovery. gorodenkoff\/iStock via Getty Images\u0026nbsp;\u003C\/p\u003E","created":"1776442339","gmt_created":"2026-04-17 16:12:19","changed":"1776442339","gmt_changed":"2026-04-17 16:12:19","alt":" AI and machine learning provide new tools for scientists to think about drug discovery. gorodenkoff\/iStock via Getty Images ","file":{"fid":"264222","name":"file-20260129-62-3xayw4-copy.jpg","image_path":"\/sites\/default\/files\/2026\/04\/17\/file-20260129-62-3xayw4-copy.jpg","image_full_path":"http:\/\/hg.gatech.edu\/\/sites\/default\/files\/2026\/04\/17\/file-20260129-62-3xayw4-copy.jpg","mime":"image\/jpeg","size":2111750,"path_740":"http:\/\/hg.gatech.edu\/sites\/default\/files\/styles\/740xx_scale\/public\/2026\/04\/17\/file-20260129-62-3xayw4-copy.jpg?itok=h8utD5AH"}}},"media_ids":["679992"],"related_links":[{"url":"https:\/\/theconversation.com\/ai-is-reengineering-drug-discovery-by-speeding-up-testing-and-scanning-petabytes-of-data-for-connections-between-diseases-274693","title":"Read This Article on The Conversation"}],"groups":[{"id":"1278","name":"College of Sciences"},{"id":"660399","name":"College of Sciences"},{"id":"658168","name":"Experts"},{"id":"1214","name":"News Room"},{"id":"1188","name":"Research Horizons"},{"id":"1275","name":"School of Biological Sciences"}],"categories":[],"keywords":[{"id":"187915","name":"go-researchnews"},{"id":"194974","name":"go-theconversation"}],"core_research_areas":[],"news_room_topics":[{"id":"71891","name":"Health and Medicine"}],"event_categories":[],"invited_audience":[],"affiliations":[],"classification":[],"areas_of_expertise":[],"news_and_recent_appearances":[],"phone":[],"contact":[{"value":"\u003Cdiv\u003E\u003Ch5\u003EAuthors:\u003C\/h5\u003E\u003Cp\u003E\u003Ca href=\u0022https:\/\/theconversation.com\/profiles\/jeffrey-skolnick-2581183\u0022\u003EJeffrey Skolnick\u003C\/a\u003E, Regents\u0027 Professor; Mary and Maisie Gibson Chair, and GRA Eminent Scholar in Computational Systems Biology, \u003Ca href=\u0022https:\/\/theconversation.com\/institutions\/georgia-institute-of-technology-1310\u0022\u003EGeorgia Institute of Technology\u003C\/a\u003E \u0026nbsp;\u003C\/p\u003E\u003Cp\u003E\u003Ca href=\u0022https:\/\/theconversation.com\/profiles\/benjamin-p-brown-2581181\u0022\u003EBenjamin P. Brown\u003C\/a\u003E, Assistant Professor, Department of Pharmacology, \u003Ca href=\u0022https:\/\/theconversation.com\/institutions\/vanderbilt-university-1293\u0022\u003EVanderbilt University\u003C\/a\u003E\u003C\/p\u003E\u003Ch5\u003EMedia Contact:\u003C\/h5\u003E\u003Cp\u003EShelley Wunder-Smith\u003Cbr\u003E\u003Ca href=\u0022mailto:shelley.wunder-smith@research.gatech.edu\u0022\u003E\u003Cstrong\u003Eshelley.wunder-smith@research.gatech.edu\u003C\/strong\u003E\u003C\/a\u003E\u003C\/p\u003E\u003C\/div\u003E","format":"limited_html"}],"email":[],"slides":[],"orientation":[],"userdata":""}}}