Designed full-length IgG antibodies against 22 targets, including GPCR agonists
Pythia Labs reports that its all-atom generative model designed full-length human IgG antibodies for 22 therapeutic targets, and that screening at most 48 candidates per target after a single design round found single-digit nanomolar binders, including both agonists and antagonists of G protein-coupled receptors.
- Lab
- Pythia Labs
- Model
- Pythia-1, an all-atom generative model over sequence and structure jointly
- Field
- Biology
- Date
- 2026-10-26
- Human collaborators
- Robin A. Aglietti, Todd O. Yeates
Sources
Original work
What was found
Therapeutic antibody discovery is slow, gives little control over how an antibody engages its target, and has struggled with whole target classes, G protein-coupled receptors above all, where the binding site is in the membrane and the conformational state matters. The pipeline generates candidates directly in full-length human IgG format rather than as fragments, conditioning on human germline frameworks and applying sequence constraints meant to avoid known developability liabilities, and reasons over sequence and atomic structure together under explicit design constraints, which is what the paper credits for being able to aim at a specified epitope and conformational state. Reported results are single-digit nanomolar affinities for soluble proteins and single-digit nanomolar cell-binding EC50 values for GPCRs, with single-digit nanomolar inhibitory potency in functional assays for several targets, and biophysical profiles the paper calls comparable to or better than clinical-stage controls. The lead sequences are published so others can test them.
Novelty check
The claim the paper makes for itself is the first demonstration of both GPCR agonism and antagonism by de novo designed full-length IgG antibodies, which is narrower than de novo antibody design in general. This registry carries two neighbouring antibody entries, 2026-10-27-ige-dislodging-antibody-designs and 2026-10-27-bighat-antibody-fragment-optimisation, neither of which is full-length IgG design and neither of which involves GPCRs; a search of data/entries.json for Pythia, IgG and GPCR returned nothing. The first-of-its-kind claim has not been checked against the antibody-design literature here and should be before the entry is cited for it.
Caveats and known objections
Enters at claimed under the lab-announcement rule: this is a preprint from the company that makes the model, the assays are its own, and no independent party has reproduced any of it. Peer review has not happened. The screening funnel is reported as at most 48 candidates per target, but the paper does not make clear whether single-digit nanomolar hits were found for all 22 targets or a subset, so the headline reads stronger than what is demonstrated target by target. All results are in vitro, with no animal or clinical data. Every author is a Pythia Labs employee. No AI disclosure statement appears as such; the model's role is read off the method description naming Pythia-1 as the generator. Autonomy is search-scaffold rather than ai-led, the weaker reading: the generative model produced candidates inside a human-designed pipeline with human-chosen targets, constraints and assays.
No independent check is recorded on this entry yet.
Reading the primary source closely enough to say whether it supports the claim counts as a check, and you are credited on the entry.
Or on GitHub: submit a check challenge the grade send a correction or send a pull request
Flag this for triage
Signals order the review queue and nothing else. They are never published, and they never move a grade: that takes a citation.
Entry history (1 event)
- AddedEntered the registry graded Claimed and Search scaffold.
Entries are never deleted. A grade that does not hold up is downgraded on the record, with the reason beside it.
Graded claimed for verification and search scaffold for autonomy. What these mean.
Cite this entry
whataifound.org. (2026). Designed full-length IgG antibodies against 22 targets, including GPCR agonists. whataifound.org: A Registry of AI Scientific and Mathematical Discoveries. https://whataifound.org/finding/2026-10-26-pythia-full-length-antibodies
BibTeX
@misc{whataifound-pythialabs-2026-antibodies,
title = {Designed full-length IgG antibodies against 22 targets, including GPCR agonists},
author = {{whataifound.org}},
year = {2026},
howpublished = {whataifound.org: A Registry of AI Scientific and Mathematical Discoveries},
note = {Result by Pythia Labs. Verification: Claimed. Autonomy: Search scaffold.},
url = {https://whataifound.org/finding/2026-10-26-pythia-full-length-antibodies}
}