Drug design asks us to find possibility within almost unimaginable scientific complexity.
The Isomorphic Labs Drug Design Engine (IsoDDE) is our answer, bringing together multiple predictive and generative AI models to reach further across biology, types of medicine and disease areas – at the scale our mission demands.


There are estimated to be around 1060 possible drug-like molecules – far more than scientists could ever make and test. We believe answers to many unsolved diseases lie somewhere within that immense chemical space. With AI, we have new ways to guide the search.
Rather than screening every possibility, IsoDDE can generate promising designs, predict their properties, and refine them computationally before scientists decide what to make and test in the lab.

IsoDDE is designed to generalise across novel biological and chemical systems, therapeutic areas and drug modalities. One engine, applied again and again.
“A model that generalises opens up entirely new possibilities: new chemistry, new targets, new drug design opportunities.”
Kathryn Tunyasuvunakool, AI Research Scientist
A technical paper on Iso’s drug design engine reports state-of-the-art performance across three core problems in AI-driven drug discovery:
“Novel systems are where you really test generalisation. Drug discovery constantly takes you beyond what is already known, so we need models that remain useful when the biology or chemistry is unfamiliar.”
Adrian Stecula, Research Scientist
IsoDDE brings machine learning and drug discovery together from the outset – drawing on diverse data and the large-scale computing power needed to learn from it.
It’s built for the whole process: from designing a molecule to developing a potential medicine, with each step informing the next.
Our scientists use IsoDDE across a growing portfolio of internal and partnered programs, moving promising ideas toward potential medicines.

