Raphael Townshend on The Power of Small Molecule Drugs

The Harry Glorikian Show

There have been a lot of stories in the news over the last few months about AI chatbots like ChatGPT that can respond to your questions with convincing and well-written answers. These so-called large language models can tell you how to build a treehouse, how to bake a cake, or how to sleep better. But notice that word large. Behind the scenes, these models have learned which word tend to cluster together by sifting through hundreds of billions of pieces of data—basically the entire Internet, in the cast of ChatGPT, including all of Wikipedia and thousands of published books. Now imagine that another chatbot came along that could learn how to generate convincing text response by studying only, say, 18 sentences. Something like that is what this week’s guest Raphael Townshend, the founder and CEO of Atomic AI, has accomplished when it comes to predicting the structure of RNA molecules.

RNA has been in the news a lot lately too. That's in part because some of the vaccines that helped us beat back the coronavirus pandemic were made from messenger RNA, a form of the molecule that instructs cells how to build proteins (in that case, antibodies to the virus). But RNA has many other functions in the body, and if we knew how to design small-molecule drugs to attach to binding pockets on any given RNA to interrupt or modulate its functions, it could open up a whole new realm of medical treatments. The problem is, if all you know about an RNA molecule is its nucleotide sequence, it’s very hard to predict where those binding pockets might be and what kind of drug might fit into them.

As a PhD student at Stanford, Townshend designed a deep learning model to tackle that problem. The model, called ARES, started with a proposed structure for an RNA molecule with a known nucleotide sequence, and predict how that proposal would compare to real-world data. ARES turned out to be stunningly accurate, and it acquired its skills by studying a remarkably small training set: just 18 examples of RNAs with known structures. So in a way, it was using the power of small data, together with a bit of physics. Now Atomic AI is building on that original model to create an engine for discovering new small-molecule drugs that could potentially interrupt any disease where RNA is a player.

For a full transcript of this episode, please visit our episode page at http://www.glorikian.com/podcast

Please rate and review The Harry Glorikian Show on Apple Podcasts! Here's how to do that from an iPhone, iPad, or iPod touch:

1. Open the Podcasts app on your iPhone, iPad, or Mac.

2. Navigate to The Harry Glorikian Show podcast. You can find it by searching for it or selecting it from your library. Just note that you'll have to go to the series page which shows all the episodes, not just the page for a single episode.

3. Scroll down to find the subhead titled "Ratings & Reviews."

4. Under one of the highlighted reviews, select "Write a Review."

5. Next, select a star rating at the top — you have the option of choosing between one and five stars.

6. Using the text box at the top, write a title for your review. Then, in the lower text box, write your review. Your review can be up to 300 words long.

7. Once you've finished, select "Send" or "Save" in the top-right corner.

8. If you've never left a podcast review before, enter a nickname. Your nickname will be displayed next to any reviews you leave from here on out.

9. After selecting a nickname, tap OK. Your review may not be immediately visible.

That's it! Thanks so much.