Patrick Kidger

⚡️ My PhD thesis is on arXiv! ⚡️ To quote my examiners it is "the textbook of neural differential equations" - across ordinary/controlled/stochastic diffeqs. w/ unpublished material: - generalised adjoint methods - symbolic regression - + more! arxiv.org/abs/2202.02435 v🧵 1/n

Alright chaps, I am in Boston for PEGS, here to chat all things AI + antibody engineering. Send me a message if you're around! 🧬 (PS I'm hiring! Pharma x AI startup, come make drugs with us.)

@thisismadani Congrats!

AI has two modes in drug discovery. Accelerate: moving faster through the existing playbook. Unlock: opening frontiers that weren't possible before. Excited to announce Profluent is partnering with Eli Lilly, the global pharma powerhouse, to unlock breakthrough medicines for patients. It's a big deal beyond the numbers ($2.25B + royalties): we’ll get to use our frontier AI models and foundational datasets to design proteins focused on large gene insertion, a therapeutic moonshot. Proteins govern almost everything in biology. We've built a generalizable AI platform to design all proteins. Onward!

@DdelAlamo +100 to this. Every now and again I get pitched by someone looking to e.g. package RNAseq data and sell to pharma. This kind of stuff is... not very useful. I'd much rather have a way to efficiently run targeted experiments (e.g. 96-well plate) than 10^n screening experiments.

I don't get this obsession with mindlessly scaling data. Pharma has tons of data. Most of it is useless. Collected on old instruments running outdated project-specific workflows and stored in long-abandoned Sharepoints & USB drives b/c emails get auto-deleted every 3 months

Yeah. :( Our instincts are the same, but again, almost no-one else in biotech approaches things this way. Some options we're thinking about: 1) bet on oneself – we'll be faster – release anyway! 2) release but with e.g. just a 1 year delay; 3) wait until later (patent, ...) but include all kinds of interesting intermediate details then; 4) ...just acknowledge that we're doing something unusual and that there aren't that many folks who'd really benefit from the data anyway. FWIW techbio is coming from a totally different culture + change is being driven by drug timelines compressing across the industry. I might just be hopelessly naive, but I think things have the possibility to improve here. :)

I absolutely hate the state of having some incredible data that we can’t publish. Like the options are to publish some random press release with no details or to publish and watch 100s of Chinese companies pursue the same, or publish with 3-4 year lag. All of these are bad. In general, my instinct is to share important data with the world (a rising tide lifts all boats etc), but I know that basically no other company thinks this way.

We're growing the AI team at @ManifoldBio, starting with a role to train protein foundation models on our proprietary data. I believe Manifold is the most interesting place to work on protein design. We're designing and testing millions of binders per month, including in vivo, and accelerating. No one else has data like this. If you have deep experience pretraining or fine-tuning protein models and want to work somewhere the data actually lets you push beyond what public datasets can enable, please reach out.

Who's the most talented software engineer you know that should stop building SaaS products and instead join a company that builds in the physical world? Refer them to me and get a 2000 USD referral bonus if we end up hiring them!

🚀 Exponax v0.2.0 — fast & differentiable PDE solvers in JAX New: 3D Navier-Stokes on a single GPU, wave equation stepper, improved dealiasing & memory efficiency 4096² / 256³ on 24GB consumer GPUs 10k² / 512³ on A100/H100 📦 pip install exponax github.com/Ceyron/exponax

@Patrick18287926 @GabriCorso > probably failed in animals Seems like a reasonable guess to me!

@PatrickKidger @GabriCorso You don't need better structure prediction models to make drugs. I heard they were preparing for trials a year ago - probably failed in animals?

So I have mixed feelings about IsoDDE. It's an AF4, it's much better on hard problems, and I don't want to understate their technical achievement. But also, it's been five years since founding, and success is measured in drugs, not models. Where are the drugs? 1/

@BiomedErs For a phase 1 then 5 years should now be pretty easy, at least in antibodies. For a famous reference then see China doing it in way less than that; the west is also still doing pretty well here.

@PatrickKidger isnt 5 years short for drugs? I assume it takes a bit to develop the model, and then ages for the drugs?

Excited to announce our next In Silico event on February 25th, 6pm at @localglobevc in King's Cross! Don't miss great talks from @McclainThiel from UCL , @PatrickKidger from @cradlebio, @kpetrovvic from Oxford, and Ivan, PD at @ARIA_research. Reg here: luma.com/gb3uso7t

Heck yeah, this is awesome :D Integrated into everyone's favourite de novo tooling (Boltz, ...) when?

Heck yeah! :D

@owl_posting @leashbio The 'Money Stuff'-style introduction was excellent, I'm a big fan. (Though I don't know if that was an explicit inspiration or not!)

An ML drug discovery startup trying really, really hard to not cheat owlposting.com/p/an-ml-drug-d… on the 12-person, Utah-based startup @leashbio, their culture of rigor, and the many ways small molecule models accidentally learn the wrong thing

@owl_posting Huge +1 from me in favour of the podcast format. - Stronger guarantee for the reader that the piece is technically accurate. - It's great market research to know the in-depth view of a particular company / research groups.

though i have a few more episodes planned, i am kinda unsure whether the podcast format will survive long term its fun, i learn a lot from the prep work i have to do for each episode, and it does definitely encourage me to explore fields i never would’ve otherwise but it is difficult to shake the feeling that the marginal value of the 20~ hours of work that goes into each one of these things would be *much* better spent on a very long essay about the same subject. you can cover everything, you can discuss more controversial things, and its fun to write things on the other hand, there is an aura of humanization of biology that comes from having a real life person speak about their field, and i like that a lot tough balance when i first started the podcast, i approached it from the lens of ‘this is a vehicle for me to spread information that i could not, in a million years, communicate myself’, but maybe i should just try to do it myself, get peoples takes on what i get wrong, and iterate one of the upcoming essays is an attempt to see how well i can communicate something that, a year back, i wouldnt have trusted myself to do. curious how that one does. it was a lot of hitting myself over the head to finish it, and it took months, but it also felt so much more rewarding there is no conclusion here, we live and we learn

we have nine roles open @ Loyal right now - come be a part of the final sprint to bring to market the first FDA-approved longevity drug.

Interesting post from @owl_posting asking whats next for antibody design. I agree with him that in vivo properties are likely the answer, but the question is how do we get there? The PDB has enabled de novo binding models to be incredibly successful. Properties like developability likely are implicit in the models, given the data they are trained on is well formed antibody crystal structures. Fwiw at @ManifoldBio using our open source mBER de novo design approach, we also see very similar affinity/developability properties for our molecules, when we get around to it we might update the preprint to reflect this. Of course you are welcome to try out mBER as it is open source: github.com/manifoldbio/mb… Back to in vivo properties, the challenge here is that unlike the PDB, we don't have an accurate and high throughput dataset for antibodies in terms of PK/PD or ADAs etc. Using standard methods, these traits are quite hard to measure in high throughput, which make learning them very challenging. And arguably, these properties are more important than binding for making a successful drug. These properties are complex, essentially you are asking "I have a binder that I know binds a target of interest in vitro: but does it get to the right place, bind the right target (and not the 20k other possibilities in the body), last long enough to enact its function (i.e. PK) and then perform the function on the target (inhib, activate etc) in a highly complex living system that is nothing like the petri dish that the molecule likely was initially tested in before?" To some extent, binding has been a solved problem for quite a while. If you talk to folks like Dane Wittrup (inventor of yeast display, founder of Adimab), he will tell you that using yeast display Adimab can design binders to specific epitopes / gpcrs, whatever you want. They will bind with high affinity and specificity in vitro. These new de novo methods indeed speed this process up by a couple months, but fundamentally the really question is still, does your molecule work in a living system, and have you optimized for all the properties (PK,PD, ADA) that will enable clinical success. This is exactly what why we built @ManifoldBio. We saw the need for high throughput in vivo data to unlock the real power of AI. AI / de novo design are only as good as the data they are trained on, without in vivo data, we will never learn in vivo properties! So we built a measurement engine to solve that. We have generated PK data on over 12,000 molecules to over 100 targets of interest to date. We have generated in vivo tissue enrichment data on half a million molecules to almost a thousand targets. This is the data we believe will unlock the true promise of AI. Lot's of folks talking about virtual cells, but perhaps we should start thinking about virtual organisms. Even if you had a virtual PK model, this would be a huge benefit to drug discovery. In fact, we already are building such a model, more details soon :)

Next company to release cool protein designs on @proteinbase! Check out @cradlebio's competition winning EGFR binders. Many more data drops on the way 👀

We’re excited to share a new multi-year collaboration with @TakedaPharma, building on the success of our first engagement. Under the agreement, Nabla will receive double-digit millions in upfront and research payments and is eligible for success-based payments exceeding $1 billion. The partnership deploys Nabla’s AI-driven JAM platform across Takeda’s early-stage programs to include de novo design of antibodies in parallel for multiple targets, multispecifics, challenging targets, and other custom therapeutics. Read more below