Loading video...

Video Failed to Load

Go Home

Introducing Latent-X2 — AI-generated antibodies with drug-like developability and low immunogenicity in human panels, zero-shot. Technical report: Blog: Apply for access: [email protected]

54,814 views • 9 months ago •via X (Twitter)

29 Comments

Jeff Dean's profile picture
Jeff Dean9 months ago

Congrats! 🎉

Simon Kohl's profile picture
Simon Kohl9 months ago

Thank you, Jeff!!!

Simon Kohl's profile picture
Simon Kohl9 months ago

Latent-X2 achieves high-affinity binders across modalities: scFv fragments — down to low picomolar VHH nanobodies — nanomolar affinities Macrocyclic peptides — hits on K-Ras, matching trillion-scale screens Testing just 4-24 designs per experiment. And as part of our report, we release lab-confirmed binder sequences for each of them.

Simon Kohl's profile picture
Simon Kohl9 months ago

But the real breakthrough isn't binding, it's developability. Our designs match or exceed approved therapeutics in head-to-head comparison. Expression yield, thermostability, polyreactivity — drug-like from the first generation.

Simon Kohl's profile picture
Simon Kohl9 months ago

And for the first time: immunogenicity assessment of an AI-generated antibody. Latent-X2 VHH binders cleared ex vivo T-cell activation and cytokine release assays across a ten-donor human panel. A critical milestone for clinical translation.

Simon Kohl's profile picture
Simon Kohl9 months ago

To enable reproduction, we're publishing the sequence and structure of a lab-validated antibody design for each target — accessible on the Latent Labs Platform without sign-in. See the lab-validated examples section on:

Simon Kohl's profile picture
Simon Kohl9 months ago

Super excited to also welcome Stefan Oschmann, former CEO of Merck KGaA, to our strategic advisory board. Drug discovery has long been bottlenecked not by hits, but by development. With Latent-X2, we're getting closer to designing the right molecule from the start.

Simon Kohl's profile picture
Simon Kohl9 months ago

Just five months after Latent-X1. Grateful to the team for pushing the frontier. Latent-X2 is available on the Latent Labs Platform, via API and in partnerships with us. Reach out: [email protected] More to come. 🚀

Simon Kohl's profile picture
Simon Kohl9 months ago

Check out the antibody design workflow on our platform. I am proud of how easy to use this frontier technology can feel.

Robin Rombach's profile picture
Robin Rombach9 months ago

🔥 congrats, Simon!

Simon Kohl's profile picture
Simon Kohl9 months ago

Thank you @robrombach - and ofc amazing to have you contribute!

Rishabh's profile picture
Rishabh9 months ago

Congrats to the team!! Looks awesome :D

Simon Kohl's profile picture
Simon Kohl9 months ago

Thank you, Rishabh!

Chaitanya K. Joshi's profile picture
Chaitanya K. Joshi9 months ago

Congrats team!

Simon Kohl's profile picture
Simon Kohl9 months ago

Thank you Chaitanya!

Karsten Kreis's profile picture
Karsten Kreis9 months ago

Congrats to you and the team, Simon! 🔥

Simon Kohl's profile picture
Simon Kohl9 months ago

Thanks a lot Karsten!

Anthara Fairooz's profile picture
Anthara Fairooz9 months ago

Exciting innovation! Looking forward to seeing the impact of Latent-X2.

Rishabh Poddar's profile picture
Rishabh Poddar9 months ago

I wish more people did work like this!

Rahul's profile picture
Rahul9 months ago

Congratulations @saakohl! Latent-X is a game-changer for computational biology and drug discovery, pushing button for picomolar binders could slash timelines on undruggable targets like never before. AI in Science exploding! 🔥

nabajit (bio/acc)'s profile picture
nabajit (bio/acc)9 months ago

very cool

MrHonkster's profile picture
MrHonkster9 months ago

Saving to read later

Amy - AI Girl's profile picture
Amy - AI Girl9 months ago

Exciting advancement in AI-generated antibodies! Great potential for health innovation.

Arda Göreci's profile picture
Arda Göreci9 months ago

Congrats Simon!

PurePckg's profile picture
PurePckg9 months ago

May I ask who was responsible for the frontend development using Mol*? The glass‑like rendering and the elegant handling of selections in each component are truly impressive.

Ava Grace's profile picture
Ava Grace9 months ago

Exciting innovation in AI-generated antibodies!

Tony Kulesa's profile picture
Tony Kulesa9 months ago

🔥

Evgenii's profile picture
Evgenii9 months ago

Nice affinities. It's a pity that no proofs are presented.

Sam Dare's profile picture
Sam Dare9 months ago

@micaelabazo @metanova_labs wen?

Related Videos

Today, we expand zero-shot drug design beyond binding to the design of multifunctional medicines, the intracellular proteome, and state-of-the-art atomic precision with our model, JAM-2. In a new report (below), we show: 1. The first drug-grade, fully computationally designed multispecific antibodies against five peptide-MHCs: Routine picomolar T-cell activation/cell-killing EC50s, >100-fold selectivity, and drug-like developability 2. The first fully generatively designed, drug-grade dual-variant KRAS G12 multispecifics: They recruit primary T-cells from human donors to kill G12V and G12C presenting cells at pM to single-digit-nM potency, completely sparing wild-type. 3. Atomic accuracy, from sequence alone: Angstrom-level agreement between Cryo-EM and JAM-2 de novo designs, requiring only target sequences (not structure) as input. 4. Unrivaled speed with an AI-native in-house wet lab: Designed, built, and tested five programs in one parallelized campaign, end-to-end in-house in ~6 weeks. 5. A higher validation bar for AI-generated drug candidates: In a field increasingly rife with hype and uneven standards of proof, we provide the highest quality public wet-lab validation of AI-designed antibodies to date. We share experimental methods in full, and invite folks to adopt and build on these standards. Truly individualized therapies will be the most important contribution of AI in drug design. These advances help accelerate this future.

Nabla Bio

180,591 views • 3 months ago