Most pharmaceutical companies use software to model chemical interactions, with the hope of speeding up the drug development process. But it’s typically a small component of a complex array of approaches. Nimbus Discovery, a startup based in Cambridge, Massachusetts, is using computational chemistry to drive the entire process.
The company emerged from a partnership with Schrödinger, a maker of computational drug discovery software, and venture capital firm Atlas Venture. Nimbus will use Schrödinger’s software, computing power, and modeling experts to develop drugs for disease-linked proteins that have historically been difficult to target.
If successful, this computationally driven approach could make drug development faster and cheaper by making much of the trial and error process virtual. Nimbus recently raised $24 million in venture funding. Bill Gates was one of the investors.
Schrödinger’s software, which is used by many pharmaceutical companies, models the various chemical forces that drive a candidate drug molecule to bind to a specific spot on the target protein. That allows drug developers to predict how well various candidate molecules bind to targets of interest. While this approach has been in use for about two decades, it has yet to truly transform the drug-discovery process.
Nimbus researchers think that part of the reason is that most tools fail to incorporate the thermodynamics of the resident water molecules in the protein’s binding site. “The need for improved water models is a widely acknowledged yet seldom-addressed limitation of current methods,” says Christopher Snow, a postdoctoral researcher at Caltech who is not involved with the company. It’s difficult to model the energy of water molecules.
WaterMap, a new tool from Schrödinger that predicts how water will affect the binding reaction, could overcome that barrier. “We think we can use our technology to transform the way drug development is done,” says Ramy Farid, president of Schrödinger and cofounder of Nimbus. Researchers have used WaterMap to explain the success or failure of some molecules, as well as to develop new candidate molecules. “It led in a number of cases to rapid development of drug candidates that were of higher quality than what appeared to be otherwise possible,” says Farid.
The startup spent its first year using the software to narrow down a list of 1,200 potential drug targets, chemical binding sites on different disease-linked proteins, to a list of 20 that looked most amenable to the technology. (That depended on a number of factors, including knowledge of the protein’s three-dimensional structure, its desirability as a target for disease, as well as the number of water molecules that reside in the binding site.) The company will focus on targets involved in inflammation, oncology, metabolic disease, and antibiotics.
The most advanced target to date is called IRAK4, a kinase enzyme that plays a role in inflammation and drives an aggressive form of non-Hodgkin’s lymphoma. Researchers conducted a virtual drug screen, looking for molecules that would bind to IRAK4, and then put those virtual molecules to the test by synthesizing them and running real chemical reactions. “We have been able to quickly find a highly selective molecule with drug-like properties,” says Rosana Kapeller, Nimbus’s chief scientific officer. It took just nine months to go from virtual screening to testing in animal models of disease.
“We have seen powerful examples of how minor changes to the molecule can result in profound changes in binding,” says Bruce Booth, one of Nimbus’s cofounders. By displacing one “unhappy” molecule, a high-energy water molecule in the binding site, “we can improve binding a hundred-fold,” he says.
While WaterMap is available to pharmaceutical companies for purchase, Farid says, the newness of the technology, and the fact that it requires intense computing power, has made it difficult to implement effectively. Part of the reason for founding Nimbus, he says, was to demonstrate how powerful the tool can be.
But it remains to be seen how significantly the WaterMap tool will speed drug discovery or how broadly applicable it will be. It may turn out to be very useful for some targets but not others.
A quick guide to the most important AI law you’ve never heard of
The European Union is planning new legislation aimed at curbing the worst harms associated with artificial intelligence.
It will soon be easy for self-driving cars to hide in plain sight. We shouldn’t let them.
If they ever hit our roads for real, other drivers need to know exactly what they are.
Crypto is weathering a bitter storm. Some still hold on for dear life.
When a cryptocurrency’s value is theoretical, what happens if people quit believing?
Artificial intelligence is creating a new colonial world order
An MIT Technology Review series investigates how AI is enriching a powerful few by dispossessing communities that have been dispossessed before.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.