---

TLDR: We build fast and experimentally accurate generative AI-based simulations of molecular interactions for pharma and biotech companies. These simulations can tell us whether a drug is going to bind to a protein or how quickly a drug will act once it is consumed by a patient.

---

Our Team

We are a team from the University of Cambridge with 30+ years of combined experience in AI and Molecular Modelling, including:

• Javier, who scaled probabilistic AI methods 1000x during his PhD, turned down a research scientist position in Big Tech to start Angstrom AI.
• Laurence, who developed FAB, a method for training AIs from physical equations without training data, during his PhD and was a researcher at InstaDeep before its acquisition by BioNTech in 2023.
• Miguel, Javier and Laurence's PhD supervisor, author of foundational research on AI for drug discovery and generative modeling.
• Gabor, who developed MACE, a state of the art quantum mechanically accurate force field.

Miguel and Gabor's work has been cited over 40k times.

The Problem

During drug discovery and development, pharmaceutical companies need to understand how drug molecules interact with other molecules in the human body to determine the drugs’ efficacy and safety.

Conventional methods to assess molecular interactions are unsatisfactory:

• Wet lab experiments are accurate but slow and expensive.
• Machine learning prediction methods, like AlphaFold, are fast but inaccurate. They are limited by the quality and quantity of training data, which must be generated by lab experiments.
• Molecular Dynamics simulates interactions by rolling out the equations of physics, offering a balance between wet lab and machine learning predictions in accuracy and speed. These factors depend on the model of physics used. More accurate models are more expensive to run. This makes the method compute bottlenecked rather than data bottlenecked.

Here is a video of a simulation of 64 water molecules. https://www.youtube.com/watch?v=18AuJEVnwWA

Solution: Fast and Accurate Molecular Simulations

We run molecular simulations, keeping us in the compute constrained regime, but we combine 1) quantum mechanically accurate models of physics 2) generative AI that allows us to run these models quickly.

1. We use the MACE (multi-atomic cluster expansion) physics model, which accurately reproduces quantum-mechanical interactions. It was developed by Gabor, our co-founder. In collaboration with our academic partners, we recently showed MACE simulations are the first ever to provide accuracy comparable to lab experiments when estimating the hydration-free energies (a quantity relevant to drug bioavailability). Below is a plot from the resulting publication. However, MACE is computationally expensive, each of the results from the below plot required 1 week of compute on 8 A100 GPUs.

   

2. We use diffusion models to accelerate MACE simulations, making them computationally affordable. Our models generate states consistent with physics, but the transitions between states are non-physical and significantly faster. Here is a video of our AI simulating the same water box as above.

https://www.youtube.com/watch?v=aK1MNLsQg2Y

Demos!

Here are a couple of examples to show off the type of stuff you can do with our generative models.

Speeding up Supercool Water

Supercool water—liquid water below freezing, here at -40°C—is notoriously difficult to simulate with traditional methods because cold molecules move slower. The plot below shows the correlation between water molecule orientations across simulation steps. Our AI introduces about 10,000 times more information per step compared to traditional simulations.

Hydrating Methane

Modeling interactions between molecules and water allows us to calculate how quickly the molecules will dissolve and their bioavailability as drugs. Here are videos of a traditional simulation of a methane molecule interacting with water and our AI simulation. Our AI prioritizes the movement of the methane molecule and its surrounding water, which are the parts that matter for solubility and bioavailability calculations.

https://youtu.be/EmTit8tC0TE   

https://youtu.be/pq6Al2SATlg

These are the first ever genAI accelerated, physically accurate molecular dynamics simulations incorporating the interaction of many molecules. We are scaling up - so stay tuned!

Our Ask

Reach out to info@angstrom-ai.com if

• You are in pharma or biotech and interested in learning about the theory of diffusion models and quantum-mechanically accurate models of atomic interactions. We would be happy to give a talk on our research or have a more informal chat over Zoom.
• You have friends who work in pharma or biotech and are interested in computational methods. We would love to meet them!