How GUIDE Is Revolutionizing Drug Development

We need effective drugs and vaccines to protect our soldiers from biological weapons, but developing countermeasures for each threat is a time-consuming process. To meet this challenge, the DoD is leveraging advances in AI and biocomputational models to revolutionize the drug development process. Enter GUIDE, the Generative Unconstrained Intelligent Drug Engineering program. Watch this video to learn more.



Transcript

00:00:03 we need effective drugs and vaccines to protect our soldiers from biological weapons but developing counter measures for each threat is a timeconsuming process to meet this challenge the Department of Defense is leveraging advances in artificial intelligence and biocomputational models to revolutionize the drug development process guide or the generative unconstrained intelligent

00:00:26 drug engineering program is a groundbreaking capability that will massively decrease the time required to develop medical counter measures originally conceived as a DARPA project guide has matured Under The Joint program executive office for chemical biological radiological and nuclear defense as a new program to discover antibodies and other types of medical

00:00:49 counter measures with unprecedented speed and accuracy what guids to do is to generate antibodies that are going to keep us safe so that could be the people that protect us um our our military our police our uh Medical Professional when they go into one of these regions or into an outbreak situation we need to have something that can protect them um and we need it fast one type of

00:01:11 countermeasure guide develops is antibody treatments antibodies are naturally produced upon exposure to a specific pathogen they prevent future infection and are a key part of the immune system in the virus that causes covid for example the best antibody Target is a small part of the virus's spike protein protein that it uses to attach enter and ultimately hijack a

00:01:34 human cell God uses the antibodies DNA sequence to generate a 3D representation of the antibody and from this rapidly develops new antibody drug treatments this is done through a tight iterative process of computational and biological experimentation separated into three phases codesign test and learn the guide codesign system uses AI

00:02:08 simulation and structural bioinformatics to model in 3D how an antibody or another countermeasure might bind to its intended target like a vital protein we call this phase codesign because we simultaneously consider multiple factors that are important for a drug product not only do we want the an body to bind very strongly to its Target but we also want it to bind to

00:02:35 many real or potentially emerging Varian or strains of the virus for example while still being safe to administer to a human and be robust enough to withstand the harsh processes in manufacturing at scale we do all of this upfront in the beginning simultaneously before we send the candidates off to experimentation we do this by modifying

00:02:58 a closely related antibody and and then evaluate whether the new antibody design is better or worse at binding to the desired Target but we also have to consider variants of the virus which increases the complexity the space of antibody designs is about 10 to the 40th uh that is too large of a space to fully uh synthesize in the laboratory all possible antibodies to any particular

00:03:23 antigen even by leveraging the fastest supercomputers in the world it is not possible to evaluate every comination to solve this problem we use custom computer models designed to prioritize certain changes to the antibody molecule ultimately our pipeline will prioritize just a few hundred designs predicted to meet our criteria these need to be tested or validated in a laboratory to

00:03:48 ensure they behave as expected here we leverage the guide rapid response laboratory which contains state-of-the-art high throughput testing capabilities first the antibodies computational designed genomic sequence is converted into DNA and inserted into biological cells these cells then Express the new antibodies which we quickly expose to the viral Target if

00:04:11 they bind well the antibodies are considered strong candidates for further testing at our partner Laboratories High throughput processes enable these tests to be completed in record time allowing us to iterate quickly this is accomplished by leveraging robotic Automation and highly sensitive instruments that can evaluate small quantities of antibodies once we

00:04:33 complete our test phase our machine learning models are updated with the results but there are good candidate was found or not the information gained from this process can be used to retrain our models for the next design campaign guide is one piece of a comprehensive Department of Defense drug development pipeline that includes other jpo programs and partners for example

00:04:54 promising antibodies also require further testing by drug developers to understand how drug interacts with the human body who seeking FD approval for clinical use we are a highly collaborative team and want to ensure the very best Technologies are used in our drug Discovery pipeline as such we partner with other National Laboratories government agencies leading academic

00:05:18 institutions and Industry we look forward to working with you too