Software engineer in computational chemistry (he/him)
Joined June 2019
- Tweets614
- Following957
- Followers389
- Likes2.6K
Surprisingly balanced take, welcome to reality
There’s a really pronounced divide between people who are on the AI side of things and people who are on the biology side of things. Even among people working in AI, the divide persists based on where they’re focused.
The AI side, adherents of scaling laws and the bitter lesson, believe that AI getting better means AI being more capable of novel discoveries in medicine (that go beyond discovering new drugs in well understood targets or speeding up dry lab work etc).
The biology people think the idea of a bitter lesson in biology is kind of ridiculous and that regardless of AI’s capabilities it’s going to be at least 5-10 years until you could hope to reliably simulate human toxicology in silico (right now AI/ML can’t even reliably simulate in vivo studies like binding affinities or functional assays).
Even the most AI pilled of them, who will list all the areas that AI can reduce the time spent on various areas, don’t think the concept of AI discovering novel targets or generating new understanding of causal biology is realistic in the near term. Most, in fact, will readily characterize the assumption that even an “ASI” level model could find cures hidden in our current understanding (without expanding that understanding with significantly more biological data) as patently false.
This was a big part of why we wanted to experience it ourselves by using AI for a (very) small portion of what it can do in biology. It was both mind blowing and revealed some poignant insights into how different the problem of making AI better at something like coding is to the problems of biology.
Even if it unlocked the ability to generate new molecules at a stunning pace, its simulation (in silico) of binding only had 49% accuracy. And if we wanted to improve that, we’d need to feed it back the data from the lab - data that took about 30x as long to get as it took to generate the molecules. And these were tests in a test tube (turn of phrase), imagine how long it would take if we wanted to improve the model with data that required recruiting humans and monitoring their reactions.
Human biology is very complex, and adding a drug - especially a novel drug and doubly so for a novel drug acting on a novel mechanism - means mapping interactions that have nth-order-effects that we don’t have the data to predict. Virtual cells have made advancements enabled by AI, but the perturbations equate to a small fraction of the possible.
This is before you even get into curing disease, especially uniquely human diseases. Brain function involves cellular identities that are unique to humans, meaning animal studies are of little use.
How long does it take to understand what the side effects of chronic administration of a drug in humans are? AI can speed up analyzing the data, or designing the trial or even getting it approved…but it can’t make a 6 month trial complete in 3 months.
I think its important to point out that significant advancements can be made in AI that contribute to biology, while also being realistic about how much it speeds up something as complex as curing a specific cancer or diseases of the brain.
Hugo MacDermott-Opeskin retweeted
Retro and lessons learned on the structure track of the OpenADMET PXR challenge, where teams had to predict poses in a notoriously large, flexible and promiscuous binding pocket on >150 unpublished structures.
openadmet.ghost.io/fantastic…
Hugo MacDermott-Opeskin retweeted
OpenADMET continues to cook! Q3 updates from the team.
openadmet.ghost.io/openadmet…
Hugo MacDermott-Opeskin retweeted
Agentic models can act, observe consequences, and decide what to do next.
This is great for coding, but if you want them to do science, lab experiments take forever.
To fix this, we made a lab-in-the-loop with zero latency!
1/10
Hugo MacDermott-Opeskin retweeted
@JacobTref nails the challenge of drug disco on the head
"...we collectively don’t know enough to deliver treatments that are excellent for patients"
Looking forward to seeing what else comes out of @FoundationOAI
nature.com/articles/d41586-0…
Hugo MacDermott-Opeskin retweeted
Halfway through the OpenADMET Cyp challenge and intermediate leaderboards are up (who is "preheat-to-450"??). Not too late to enter! We are also opening up a new structure track with 20 new Cryo-EM structures of 3A4 complexed to ligands! openadmet.ghost.io/halfway-t…
Hugo MacDermott-Opeskin retweeted
Wrote an essay on how AI has been changing things at Octant, and how we've evolved from thinking about lab-in-the-loop applications to more agentic ways to work on frontier drug programs, where you don't even know what question to answer.
Hugo MacDermott-Opeskin retweeted
An OpenAI agent swarm operating under poorly airgapped research conditions accessed my CBA Mastercard and bought 4 schooners of Newtowner between 5:52pm and 7:15pm last night
Hugo MacDermott-Opeskin retweeted
1. Connected data: Following biology across multiple steps 🔗
Biology unfolds at different scales, but most life science datasets capture only one part of reality. A cell atlas may measure molecules across thousands of cells, while single-molecule imaging follows one cell only. Both views are valuable, but because such datasets are usually created independently, it can take years of follow-on work to determine how scientific results relate to each other. We want to support individual research teams to move deliberately between these scales and to create rich, connected datasets whose different layers reinforce one another.
For example, our grant to @UCSF will allow the OpenADMET team to measure key molecular properties and interactions for tens of thousands of compounds. The team will then connect this broad profiling to two views of drug transport: experimentally determined structures of transporter proteins with selected molecules bound, and assays measuring how these proteins functionally interact with the molecules. Finally, they will test a subset of these compounds in models of the human blood-brain barrier, comparing those results with in vivo animal measurements.
This approach to data collection will create an unusually complete account of what determines a drug’s movement through the body – one that enables better prediction from AI models at the level we humans care about.
Hugo MacDermott-Opeskin retweeted
Ooooh GPCR data 👀
Thanks to the @gatesfoundation for supporting us to expand into Tox for the work in OpenADMET! @OctantBio will be helping build scaled/cheap data for targets like hERG and GPCRs and excited to keep the competitions coming for the next few years!
openadmet.ghost.io/putting-t…
Super excited to be tackling new targets at OpenADMET with @gatesfoundation !
OpenADMET is breaking in to the "T" of ADMET with the @gatesfoundation , first applying our same approach to hERG & aminergic GPCRs for better mechanistic understanding and prediction. Thank you to the @gatesfoundation for your continued support!
openadmet.ghost.io/putting-t…
Hugo MacDermott-Opeskin retweeted
We’re partnering with @Anthropic to launch the biggest Protein Design Competition in the world, challenging people around the world to use AI to design new potential drug candidates for diseases that affect millions of lives.
The competition will feature five challenges, each focused on a specific disease or biological mechanism. Compared to previous competitions, it will be a big step-up in complexity and scale to push the boundaries of AI-driven protein design.
Together with Anthropic, we’re sponsoring over $1 million in experimental validation, making it possible to test more than 5,000 protein designs in our automated lab at no cost to participants. Anthropic is providing an additional $1 million in Claude credits.
All experimental results will be published openly on @Proteinbase, including designs that didn’t work, so anyone can access the data and build on what we learn.
The competition is open to everyone and free to enter. It will feature 3 tracks:
- Track 1 is aimed at expert protein designers, with up to 20 teams to be selected.
- Track 2 is targeting life science academics and industry researchers.
- Track 3 is open to everyone from tech enthusiasts to high-school students.
By combining Anthropic’s models with access to our automated lab, we want to make it possible for anyone with a laptop and an internet connection to join the global effort to advance human health with AI.
A big thanks to @Modal for contributing compute for protein design and to @TwistBioscience for contributing the DNA for the experimental validation!
Sign up link below -
Hugo MacDermott-Opeskin retweeted
Love seeing the OpenADMET data and challenges results in the wild and pushing drug discovery forward. The folks over an @inductive_bio have been incredible participants in the community, and I love that the challenges are a real testing ground for model performance. Congrats!!
Today @inductive_bio is launching Beacon-2, which can predict human dose directly from chemical structure.
A dirty secret of AI for drug discovery is that most systems optimize the wrong reward, making them near useless in practice. Human dose is the right reward – it’s what every drug program is already optimizing. Making it computable will unlock the real-world impact of agentic drug discovery.
Beacon-2 builds on our Beacon-1 models that beat 750+ competitors to win all three OpenADMET blind challenges. Whereas Beacon-1 provided a suite of models for individual molecular properties, Beacon-2 is a single system that predicts human dose end-to-end.
It has three configurable modules: an ADMET-PK foundation model, a potency module that can use our fine-tunable co-folding affinity model or physics-based methods like FEP, and a PK/PD model with physiologically-informed mechanistic models of pharmacokinetics.
We share validation results from Beacon-2 in a blog we are releasing today. We show that across 20 real-world drug programs running through Inductive, dose predictions from Beacon-2 agree with those from experimental data with a Spearman’s ρ of 0.61. And in a similar study on 325 publicly available compounds from the ExpansionRx OpenADMET competition, we show a Spearman ρ of 0.79.
As a proof-of-concept, we gave Beacon-2 to Indy, our medicinal chemistry agent, which used it over 5 autonomous design cycles on a recently disclosed SARS-CoV-2 program to cut predicted human dose 17x, which can be the difference between a drug program that gets killed and that makes it to clinical trials.
We compared the compounds that Indy designed using Beacon-2 against compounds from an optimization loop that used a potency-only reward function. The Beacon-2 compounds were favorable across three independent measures: they were preferred by chemists in blinded comparisons 87% of the time; were more druglike according to QED; and had a lower projected dose when computed by alternate methods relying on in vivo rat models and allometric scaling. Based on these encouraging results, we are now running prospective studies where compounds optimized by Beacon-2 will be validated in the lab.
Read the full post here: inductive.bio/news/beacon-2
At Inductive, our long-term goal is to build superhuman chemical intelligence, and by completing the reward function for autonomous drug discovery, Beacon-2 is an exciting step toward that goal.
Reach out if you’d like to get involved.
Hugo MacDermott-Opeskin retweeted
Today @inductive_bio is launching Beacon-2, which can predict human dose directly from chemical structure.
A dirty secret of AI for drug discovery is that most systems optimize the wrong reward, making them near useless in practice. Human dose is the right reward – it’s what every drug program is already optimizing. Making it computable will unlock the real-world impact of agentic drug discovery.
Beacon-2 builds on our Beacon-1 models that beat 750+ competitors to win all three OpenADMET blind challenges. Whereas Beacon-1 provided a suite of models for individual molecular properties, Beacon-2 is a single system that predicts human dose end-to-end.
It has three configurable modules: an ADMET-PK foundation model, a potency module that can use our fine-tunable co-folding affinity model or physics-based methods like FEP, and a PK/PD model with physiologically-informed mechanistic models of pharmacokinetics.
We share validation results from Beacon-2 in a blog we are releasing today. We show that across 20 real-world drug programs running through Inductive, dose predictions from Beacon-2 agree with those from experimental data with a Spearman’s ρ of 0.61. And in a similar study on 325 publicly available compounds from the ExpansionRx OpenADMET competition, we show a Spearman ρ of 0.79.
As a proof-of-concept, we gave Beacon-2 to Indy, our medicinal chemistry agent, which used it over 5 autonomous design cycles on a recently disclosed SARS-CoV-2 program to cut predicted human dose 17x, which can be the difference between a drug program that gets killed and that makes it to clinical trials.
We compared the compounds that Indy designed using Beacon-2 against compounds from an optimization loop that used a potency-only reward function. The Beacon-2 compounds were favorable across three independent measures: they were preferred by chemists in blinded comparisons 87% of the time; were more druglike according to QED; and had a lower projected dose when computed by alternate methods relying on in vivo rat models and allometric scaling. Based on these encouraging results, we are now running prospective studies where compounds optimized by Beacon-2 will be validated in the lab.
Read the full post here: inductive.bio/news/beacon-2
At Inductive, our long-term goal is to build superhuman chemical intelligence, and by completing the reward function for autonomous drug discovery, Beacon-2 is an exciting step toward that goal.
Reach out if you’d like to get involved.
Hugo MacDermott-Opeskin retweeted
I love how the OpenADMET team is not only generating open datasets and running blind challenges, but is also sharing their methods, learnings, infrastructure and playbooks in real-time so the community can benefit from and build upon the work. Bravo 👏
Interest is growing in starting blind challenges to galvanize communities to solve important commons problems and assess frontier AI. Through our experience with OpenADMET, we've learned lessons, built infrastructure, and hope to help other communities do the same. We're building playbooks & infrastructure for others who want to start their own challenges. We will continue to share more over time. Thank you to @beckypferdehirt and Radial / @AsteraInstitute for the support (and push) to make this happen. openadmet.ghost.io/open-sour…
Hugo MacDermott-Opeskin retweeted
I've been finding that AI in drug discovery pieces these days don't really explain why I'm so excited about it. I realized that a big reason for the disconnect is b/c drug discovery is very different depending on the problems you are trying to tackle. The programs that inspire me, and what we try to focus Octant on, are programs like RevMed's RAS drug, Vertex's CF franchise, and other programs where the target was always clear; we just had to invent new tech to understand the biology and chemistry, and in the process open up new ways to drug other proteins.
These programs are arduous, and I do think there are real ways that platforms and AI can make that process easier, but I thought I'd first write about why these kinds of programs (I'm calling Frontier Drugs, for lack of a better name) are so difficult, and what they look like when they work.
Hugo MacDermott-Opeskin retweeted
Today we’re excited to launch Public Data for Health, with more than $125M in initial grants to nonprofits and universities creating and opening up high-quality scientific datasets.
AI is getting dramatically better at analyzing biology, but accelerating medical discovery depends on researchers having high-quality data to work with.
Our goal is to make more valuable scientific data broadly available so researchers can build on each other’s work and accelerate discoveries that help prevent and treat disease.
Learn more: openaifoundation.org/news/pu…
Hugo MacDermott-Opeskin retweeted
Excited for OpenADMET to partner with the @FoundationOAI !
I am in awe at the OpenADMET momentum - the blind challenges, community, and generosity of our funders all point to the massive importance of this problem. Thank you for believing in the vision!
At the OpenAI Foundation, we believe breakthroughs against disease will come from pairing intelligence with more observations of the world – in other words, more data.
Today, we're announcing over $125M in grants, launching our second science program:
openaifoundation.org/news/pu…