When scientists develop new molecules—whether for medicine, agriculture, or species control—it’s critical to know precisely what those molecules interact with. Understanding both intended and unintended molecular targets is essential for safety and effectiveness.
Now, inspired by a toxin found in cone snail venom, researchers at the Weizmann Institute of Science in Israel have created a powerful new AI-driven method to identify molecular targets, potentially revolutionizing ecological research and drug development.
This novel approach, which combines artificial intelligence with traditional research techniques, will be presented at the 69th Biophysical Society Annual Meeting in Los Angeles in February 2025.
Unlocking the Mystery of Cone Snail Toxins
The research was led by scientists Dr. Izhar Karbat and Dr. Eitan Reuveny, who sought to understand how a specific cone snail toxin, Conkunitzin-S1 (Cs1), affects fish, a common prey of these snails. While Cs1 is known to block potassium channels—tiny gateways that control ion flow in and out of cells—it has only been shown to impact insects like fruit flies. Its effects on fish remained a mystery.
“Three years ago, we tried our best tools at the time to find the target of the Conkunitzin toxin, and we failed because the tools were not good enough. And then came a big revolution in structural biology driven by artificial intelligence,” Karbat explained in a recent statement.
That revolution allowed them to take a fresh approach using AI-driven computational models.
AI-Powered Discovery
The team used a two-step AI approach to identify which potassium channels in fish were vulnerable to Cs1. First, they used AlphaFold, a powerful AI tool that predicts how molecules interact, to model how the toxin binds to different fish potassium channels. Then, they introduced ET3, a new AI model designed to track how water molecules behave around these channels.
The researchers focused on water molecules as the key to a potassium channel’s function lies in its selectivity filter, which determines which ions can pass through. If something disrupts this filter, the channel stops working. ET3 was trained to detect abnormalities in water movement around the selectivity filter, helping pinpoint exactly how Cs1 interferes with fish potassium channels.
By analyzing a much broader range of fish potassium channels than traditional methods would allow, the researchers finally identified the toxin’s exact targets and how it disables them.
“Using molecular dynamics and the new AI-driven structural tools, we were able to find the small subset of channels in fish which bind our toxins with high affinity and are probably the real target of the cone snail,” Karbat said.
Toxins for Medicine
Beyond solving a long-standing ecological puzzle, this research opens exciting possibilities for other fields—especially drug development. The new AI-driven method could help scientists predict how a drug interacts with different proteins, reducing the risk of unintended side effects.
“This new pipeline offers exciting opportunities and future prospects with ecological studies, to study real chemical interactions in real ecological systems,” Karbat noted.
More importantly, this technique could be applied in medicine to ensure that new drugs interact only with their intended targets.
“For example, if you develop a drug that would activate a channel in the human brain, you wouldn’t want the same drug to affect a channel in the human heart and cause a heart attack,” Karbat explained.
With their AI pipeline, Karbat and Reuveny have demonstrated that technology inspired by nature—in this case, a toxin from a venomous sea snail—can help advance both fundamental biology and modern medicine.
“The power of this pipeline is that we can concentrate on a target, or any molecule that we are interested in, and find its match,” Reuveny said.
Kenna Hughes-Castleberry is the Science Communicator at JILA (a world-leading physics research institute) and a science writer at The Debrief. Follow and connect with her on BlueSky or contact her via email at kenna@thedebrief.org