MSGID: 1:317/3 6476cd9c   
   PID: hpt/lnx 1.9.0-cur 2019-01-08   
   TID: hpt/lnx 1.9.0-cur 2019-01-08   
    Using AI to create better, more potent medicines    
    Novel framework could offer chemists greater drug options    
      
     Date:   
         May 30, 2023   
     Source:   
         Ohio State University   
     Summary:   
         While it can take years for the pharmaceutical industry to create   
         medicines capable of treating or curing human disease, a new study   
         suggests that using generative artificial intelligence could vastly   
         accelerate the drug-development process.   
      
      
         Facebook Twitter Pinterest LinkedIN Email   
      
   ==========================================================================   
   FULL STORY   
   ==========================================================================   
   While it can take years for the pharmaceutical industry to create   
   medicines capable of treating or curing human disease, a new study   
   suggests that using generative artificial intelligence could vastly   
   accelerate the drug-development process.   
      
   Today, most drug discovery is carried out by human chemists who rely   
   on their knowledge and experience to select and synthesize the right   
   molecules needed to become the safe and efficient medicines we depend   
   on. To identify the synthesis paths, scientists often employ a technique   
   called retrosynthesis -- a method for creating potential drugs by working   
   backward from the wanted molecules and searching for chemical reactions   
   to make them.   
      
   Yet because sifting through millions of potential chemical reactions can   
   be an extremely challenging and time-consuming endeavor, researchers at   
   The Ohio State University have created an AI framework called G2Retro to   
   automatically generate reactions for any given molecule. The new study   
   showed that compared to current manual-planning methods, the framework   
   was able to cover an enormous range of possible chemical reactions as   
   well as accurately and quickly discern which reactions might work best   
   to create a given drug molecule.   
      
   "Using AI for things critical to saving human lives, such as medicine, is   
   what we really want to focus on," said Xia Ning, lead author of the study   
   and an associate professor of computer science and engineering at Ohio   
   State. "Our aim was to use AI to accelerate the drug design process, and   
   we found that it not only saves researchers time and money but provides   
   drug candidates that may have much better properties than any molecules   
   that exist in nature."  This study builds on previous research of Ning's   
   where her team developed a method named Modof that was able to generate   
   molecule structures that exhibited desired properties better than any   
   existing molecules. "Now the question becomes how to make such generated   
   molecules, and that is where this new study shines," said Ning, also an   
   associate professor of biomedical informatics in the College of Medicine.   
      
   The study was published today in the journal Communications Chemistry.   
      
   Ning's team trained G2Retro on a dataset that contains 40,000 chemical   
   reactions collected between 1976 and 2016. The framework "learns" from   
   graph- based representations of given molecules, and uses deep neural   
   networks to generate possible reactant structures that could be used to   
   synthesize them.   
      
   Its generative power is so impressive that, according to Ning, once   
   given a molecule, G2Retro could come up with hundreds of new reaction   
   predictions in only a few minutes.   
      
   "Our generative AI method G2Retro is able to supply multiple different   
   synthesis routes and options, as well as a way to rank different options   
   for each molecule," said Ning. "This is not going to replace current   
   lab-based experiments, but it will offer more and better drug options   
   so experiments can be prioritized and focused much faster."  To further   
   test the AI's effectiveness, Ning's team conducted a case study to see   
   if G2Retro could accurately predict four newly released drugs already   
   in circulation: Mitapivat, a medication used to treat hemolytic anemia;   
   Tapinarof, which is used to treat various skin diseases; Mavacamten,   
   a drug to treat systemic heart failure; and Oteseconazole, used to   
   treat fungal infections in females. G2Retro was able to correctly   
   generate exactly the same patented synthesis routes for these medicines,   
   and provided alternative synthesis routes that are also feasible and   
   synthetically useful, Ning said.   
      
   Having such a dynamic and effective device at scientists' disposal   
   could enable the industry to manufacture stronger drugs at a quicker   
   pace -- but despite the edge AI might give scientists inside the lab,   
   Ning emphasizes the medicines G2Retro or any generative AI creates still   
   need to be validated -- a process that involves the created molecules   
   being tested in animal models and later in human trials.   
      
   "We are very excited about generative AI for medicine, and we are   
   dedicated to using AI responsibly to improve human health," said Ning.   
      
   This research was supported by Ohio State's President's Research   
   Excellence Program and the National Science Foundation. Other Ohio State   
   co-authors were Ziqi Chen, Oluwatosin Ayinde, James Fuchs and Huan Sun.   
      
       * RELATED_TOPICS   
             o Health_&_Medicine   
                   # Pharmacology # Pharmaceuticals # Alternative_Medicine   
                   # HIV_and_AIDS   
             o Computers_&_Math   
                   # Neural_Interfaces # Artificial_Intelligence #   
                   Computer_Modeling # Computer_Science   
       * RELATED_TERMS   
             o Pharmaceutical_company o Drug_discovery o   
             Artificial_intelligence o Child o Personalized_medicine o   
             Computer_vision o Pharmacology o Water_purification   
      
   ==========================================================================   
   Story Source: Materials provided by Ohio_State_University. Original   
   written by Tatyana Woodall. Note: Content may be edited for style   
   and length.   
      
      
   ==========================================================================   
   Journal Reference:   
      1. Ziqi Chen, Oluwatosin R. Ayinde, James R. Fuchs, Huan Sun, Xia Ning.   
      
         G2Retro as a two-step graph generative models for retrosynthesis   
         prediction. Communications Chemistry, 2023; 6 (1) DOI:   
         10.1038/s42004- 023-00897-3   
   ==========================================================================   
      
   Link to news story:   
   https://www.sciencedaily.com/releases/2023/05/230530174302.htm   
      
   --- up 1 year, 13 weeks, 1 day, 10 hours, 50 minutes   
    * Origin: -=> Castle Rock BBS <=- Now Husky HPT Powered! (1:317/3)   
   SEEN-BY: 15/0 106/201 114/705 123/120 153/7715 218/700 226/30 227/114   
   SEEN-BY: 229/110 112 113 307 317 400 426 428 470 664 700 291/111 292/854   
   SEEN-BY: 298/25 305/3 317/3 320/219 396/45   
   PATH: 317/3 229/426