Still, Verba’s team uses Alphafold (both 2 and 3, because they say they have different strengths) to run virtual versions of their experiments before running them in the lab. Using Alphafold’s results, they can narrow down the focus of an experiment — or decide it’s not worth doing.
It can really save time, he says: “It hasn’t really changed any of the experiences, but it’s enhanced them a little bit.”
The new wave
Alphafold was designed to be used for a variety of purposes. Now several startups and university labs are building on their success to develop a new wave of tailored drug discovery tools. This year, a collaboration between MIT researchers and the AI ​​drug company Iterative developed a model called Boltz-2, which predicts not only the structure of proteins but also How well the drug molecules will bind to their target.
Last month, startup Genesis Molecular AI released another one A structure prediction model called Perlwhich the firm claims is more accurate than Alphafold 3 for some questions that are critical to drug development. Perl is interactive, so drug developers can feed any additional data to the model to guide its predictions.
Alphafold was a big leap, but there’s more to do, says Evan Feinberg, CEO of Genesis Molecular AI: “We’re still fundamentally innovating, just with a better starting point than before.”
Genesis Molecular AI is pushing the margin of error to less than two angstroms, the de facto industry standard set by Alphafold, to less than one angstrom, 10 millionths of a millimeter, or the width of a single hydrogen atom.
“Small errors can be devastating for predicting how well a drug will actually bind to its target,” says Michael Levin, vice president of modeling and simulation at the firm. This is because chemical forces that interact at one Hengstrom may cease to do so at two. “It went from ‘they’ll never communicate’ to ‘they will,'” he says.
With so much activity in this space, how soon should we expect new types of drugs to hit the market? The jumper is practical. Predicting protein structures is just one step among many, he says: “It wasn’t the only problem in biology. It’s not like we were a protein structure away from treating any diseases.”