{"id":776012,"date":"2026-07-02T18:26:10","date_gmt":"2026-07-02T18:26:10","guid":{"rendered":"https:\/\/www.newsbeep.com\/ca\/776012\/"},"modified":"2026-07-02T18:26:10","modified_gmt":"2026-07-02T18:26:10","slug":"how-generative-ai-and-physics-can-help-design-new-antibiotics","status":"publish","type":"post","link":"https:\/\/www.newsbeep.com\/ca\/776012\/","title":{"rendered":"How generative AI and physics can help design new antibiotics"},"content":{"rendered":"<p>By 2050, scientists estimate that antibiotic-resistant infections will be associated <a href=\"https:\/\/doi.org\/10.1016\/s0140-6736(24)01867-1\" rel=\"nofollow noopener\" target=\"_blank\">with more than eight million deaths<\/a> around the world every year. <\/p>\n<p>These are bacterial infections that resist traditional antibiotics like penicillin. They can develop when you eat contaminated food, have an open wound or undergo surgery. E. coli is a good example, as <a href=\"https:\/\/doi.org\/10.1128\/microbiolspec.arba-0026-2017\" rel=\"nofollow noopener\" target=\"_blank\">several strains have become highly resistant to conventional antibiotics<\/a>. They can also arrive as secondary infections, like <a href=\"https:\/\/theconversation.com\/how-falling-vaccination-rates-are-fuelling-the-antibiotic-resistance-crisis-259682\" rel=\"nofollow noopener\" target=\"_blank\">pneumonia after a virus<\/a>.<\/p>\n<p>            <img decoding=\"async\" alt=\"A man in a white coat with a serious face at a desk\" src=\"https:\/\/www.newsbeep.com\/ca\/wp-content\/uploads\/2026\/07\/file-20260701-57-o86zhe.jpg\" class=\"native-lazy\" loading=\"lazy\"  \/><\/p>\n<p>              Professor Alexander Fleming, who first discovered penicillin, in his laboratory at St Mary\u2019s, Paddington, London (1943).<br \/>\n              <a class=\"source\" href=\"https:\/\/commons.wikimedia.org\/wiki\/File:Synthetic_Production_of_Penicillin_TR1468.jpg\" rel=\"nofollow noopener\" target=\"_blank\">(Wikimedia Commons)<\/a>, <a class=\"license\" href=\"http:\/\/creativecommons.org\/licenses\/by\/4.0\/\" rel=\"nofollow noopener\" target=\"_blank\">CC BY<\/a><\/p>\n<p>We need new antibiotics and designing them is difficult. <\/p>\n<p>It can take 10 years and more than one billion dollars to bring <a href=\"https:\/\/doi.org\/10.1038\/s41591-024-03218-w\" rel=\"nofollow noopener\" target=\"_blank\">just one new drug to market<\/a>. And 10 out of 13 new antibiotics developed since 2017 are already <a href=\"https:\/\/www.who.int\/publications\/i\/item\/9789240094000\" rel=\"nofollow noopener\" target=\"_blank\">ineffective against at least one type of bacteria<\/a>. <\/p>\n<p>A potential solution is to use generative AI models, guided by trained scientists, to come up with <a href=\"https:\/\/www.news-medical.net\/health\/Halicin-The-AI-Discovered-Antibiotic-That-Fights-Superbugs.aspx\" rel=\"nofollow noopener\" target=\"_blank\">designs for never-before-seen molecules<\/a>. <\/p>\n<p>Physics-based simulations, where a computer mimics the laws of reality, can then help us figure out whether they would make good drugs in a fast and cost-efficient way. <\/p>\n<p>The peptide haystack<\/p>\n<p>All methods need a starting point. \u201cDevelop a new drug\u201d isn\u2019t a specific enough prompt. If we were looking for a needle in a haystack, we would at least need to know which haystack to look in. One good haystack for drugs, especially antibiotics, is peptides.<\/p>\n<p>Peptides are short proteins that can perform many different functions in our bodies. For example, insulin, which is widely used to treat diabetes, is a naturally occurring peptide in the body. Vancomyin is another peptide and an important antibiotic that is created in nature as a defence mechanism by bacteria that live in the soil. Both occupy a place of honour on the <a href=\"https:\/\/www.who.int\/publications\/i\/item\/B09474\" rel=\"nofollow noopener\" target=\"_blank\">WHO Model List of Essential Medicines.<\/a><\/p>\n<p>We can use AI and physics-based simulations together to design new peptides that will kill bacteria. A good AI model has two parts: one that can quickly dream up millions of new designs (the generator) and one that can recommend which design to simulate next (the recommender). <\/p>\n<p>            <img decoding=\"async\" alt=\"A high-level diagram of a workflow showing how scientists design drugs with physics and AI\" src=\"https:\/\/www.newsbeep.com\/ca\/wp-content\/uploads\/2026\/07\/file-20260616-69-81n5b5.gif\" class=\"native-lazy\" loading=\"lazy\"  \/><\/p>\n<p>              This diagram shows the workflow scientists use when designing drugs with physics and AI. The recommender generates letters specifying the different amino acids that make up a peptide.<\/p>\n<p>Training the generator<\/p>\n<p>The recommender is a bit like the YouTube algorithm that suggests videos you might want to watch next. It\u2019s useful, important and tricky to set up just right. In fact, our lab has <a href=\"https:\/\/pubs.rsc.org\/me\/article\/doi\/10.1039\/d5me00225g\/1227679\/Towards-best-practices-in-low-dimensional-semi\" rel=\"nofollow noopener\" target=\"_blank\">recently published a research paper<\/a> that examines a few different strategies for generators and recommenders.<\/p>\n<p>In our research, we tested what kinds of information are most useful to include when training the generator. While it\u2019s possible to provide a generator with almost any available information about peptides, we showed it was better to give it just a tiny bit of very relevant information than a lot of semi-relevant information.<\/p>\n<p>Why might this be important? Because in many cases we only have a little bit of relevant information: only a few peptides out of the hundreds of thousands we know of have been experimentally tested, for example, for antimicrobial properties. We also tested different strategies for the recommender, and found a way to more accurately visualize the path that it takes through peptide search space.<\/p>\n<p>Why not just use the generator by itself?  Well, like any AI-generated content, we have to validate it before trusting it.<\/p>\n<p>A deadly dance<\/p>\n<p>That\u2019s where physics comes in! Peptides perform their functions by changing their shapes. For example, the unusual drug Ziconotide is a painkiller that works by <a href=\"https:\/\/www.dustri.com\/article_response_page.html?artId=1226&amp;doi=10.5414\/CPP44478&amp;L=0\" rel=\"nofollow noopener\" target=\"_blank\">physically \u201cjamming\u201d the proteins that send pain signals in the spine<\/a>.<\/p>\n<p>Many peptide antibiotics, or \u201cantimicrobial peptides,\u201d have fluctuating shapes that depend on their proximity to the outside of a cell. <\/p>\n<p>Each copy of an antimicrobial peptide is a dancer in a choreography. One dance is performed near the cell of a mammal, while a different, deadlier dance is performed near the cell of a bacterium. This second dance can kill a microbe like E. coli by attacking and breaking apart the membrane. <\/p>\n<p>We can validate the molecules the AI recommends by peeking at the performance our peptide dancers put on near different kinds of membranes.<\/p>\n<p>            <img decoding=\"async\" alt=\"A pile of prescription pill packs, containing white, red, pink, orange and green drugs.\" src=\"https:\/\/www.newsbeep.com\/ca\/wp-content\/uploads\/2026\/07\/file-20260630-57-5ee2da.jpg\" class=\"native-lazy\" loading=\"lazy\"  \/><\/p>\n<p>              Without effective antibiotics, routine procedures such as hip replacements and C-sections can become life-threatening.<br \/>\n              (Unsplash\/Volodymyr Hryshchenko)<\/p>\n<p>Video game physics engine<\/p>\n<p>In physics-based simulations, we put many peptides near a simplified membrane, surround the whole system with a box of water and treat every atom as a soft sphere. Then, using something like a video game physics engine, we can watch how the atoms of the peptides and membrane dance as time goes on. <\/p>\n<p>Some people call this an \u201cin silico\u201d microscope, which allows us to zoom in and watch what happens at the molecular scale.<\/p>\n<p>With this molecular microscope in hand, we can validate generated peptides. If we see modelled peptides disturbing a simplified bacteria membrane, then we can say it is likely antimicrobial. If we see it disturbing a simplified red blood cell membrane, then we can say it is likely toxic. <\/p>\n<p>By doing this, we can pre-screen novel, never-before-seen peptides for non-toxic antimicrobial activity before wasting experimental effort. And we can also use the information we get from these simulations to improve our generators and our recommenders.<\/p>\n<p>That way, scientists can avoid the time-consuming process of identifying promising peptides, and devote their laboratory experiments to validating their clinical use and safety. This could mean more, cheaper drugs, exactly when we need them most.<\/p>\n","protected":false},"excerpt":{"rendered":"By 2050, scientists estimate that antibiotic-resistant infections will be associated with more than eight million deaths around the&hellip;\n","protected":false},"author":2,"featured_media":776013,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[10],"tags":[49,48,84],"class_list":["post-776012","post","type-post","status-publish","format-standard","has-post-thumbnail","category-health","tag-ca","tag-canada","tag-health"],"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/posts\/776012","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/comments?post=776012"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/posts\/776012\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/media\/776013"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/media?parent=776012"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/categories?post=776012"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/tags?post=776012"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}