Claude-generated antibody designs that dislodge receptor-bound IgE
Claude turned Divergence Labs' goals into a design plan and then, largely autonomously, generated roughly 35,000 antibody designs in about a month towards a two-armed treatment for food allergy; early testing found three designs that bound IgE and 20 that dislodged it from immune cells, three of the 20 appearing to dislodge more of it than the standard medication, and the two arms have not been combined into a single molecule.
- Model
- Not disclosed; no model name or version given in the announcement
- Field
- Biology
- Date
- 2026-10-27
Sources
Original work
Announcement
What was found
Food allergy is set off by IgE, some of it circulating free in the blood and some attached to the surface of immune cells. Standard medication neutralises the circulating IgE only. Divergence Labs is pursuing a molecule with two arms, one that binds IgE tightly and one that prises IgE off the immune cell, so that receptor-bound IgE is removed as well. Claude converted that goal into a detailed design plan and then generated roughly 35,000 antibody designs over about a month, which Divergence narrowed down and tested. The announcement reports hits for both arms, three designs that bound IgE and 20 that dislodged it, with three of the 20 appearing to dislodge more IgE than the standard medication. It gives no affinities, no assay format and no count of how many designs were tested to reach those hits. The stated next step, building one drug that combines the two arms, has not been done.
Novelty check
Searched Europe PMC on 2026-10-27. The mechanism is established prior art rather than new: actively stripping IgE from its high-affinity receptor is a known class of approach, reviewed as disruptive inhibition in "Structural and Functional Insights Into IgE Receptor Interactions and Disruptive Inhibition", Immunological Reviews, May 2025 (doi 10.1111/imr.70031). DARPin AND E2_79 AND IgE AND dissociation returned 25 hits, E2_79 being the designed ankyrin-repeat protein used in that literature to demonstrate accelerated dissociation of receptor-bound IgE, and ligelizumab AND IgE AND receptor returned 309. IgE AND "facilitated dissociation" AND FceRI returned 0, which reflects the phrasing rather than an absence of work. What is claimed new here is therefore a set of antibody designs produced by a model, and the report that three of them dislodge more IgE than the standard medication, not the idea of dislodging it. No existing registry entry covers IgE, food allergy or anti-allergy antibodies. Divergence's technical report PDF was not read for this entry, so no check was made against Divergence's own earlier designs.
Caveats and known objections
Entered at claimed under the rule that lab announcements start there. Divergence Labs ran the testing and Anthropic is its partner, so no one outside the collaboration has assessed the result. The evidence is early in vitro screening: counts of designs that bound or dislodged IgE, with no affinities, no assay description, no replicate counts and no statistics in the announcement. The comparator, described only as the standard medication, is not named in the post; the established drug for the circulating-IgE mechanism it describes is omalizumab, but the post does not say so and this entry does not assume it. The claim that three designs dislodge more IgE than the standard medication is hedged in the source itself as what they appeared to do. Nothing here is a drug: the two arms exist only as separate designs, and combining them is listed as future work. Only hit counts are given, so the hit rate across roughly 35,000 designs is unknown, which is the figure that would show how well the generation step worked. Divergence's technical report PDF is linked and was not read for this entry. Autonomy is graded search-scaffold on the strictest defensible reading even though the announcement says Claude worked largely autonomously: Divergence set the therapeutic goal and the two-arm architecture and did all the narrowing and all the testing, which is the human-designed loop that grade describes. A reading that weighted Claude's conversion of those goals into a design plan would reach ai-led. Field is biology rather than medicine because no therapeutic or clinical endpoint was reached. One commenter on the announcement described this programme as speeding up an existing clever play in food allergies rather than a new one, which is the reading the novelty check above reaches independently.
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.
Community discussion
Graded claimed for verification and search scaffold for autonomy. What these mean.
Cite this entry
whataifound.org. (2026). Claude-generated antibody designs that dislodge receptor-bound IgE. whataifound.org: A Registry of AI Scientific and Mathematical Discoveries. https://whataifound.org/finding/2026-10-27-ige-dislodging-antibody-designsBibTeX
@misc{whataifound-divergencelabs-2026-designs,
title = {Claude-generated antibody designs that dislodge receptor-bound IgE},
author = {{whataifound.org}},
year = {2026},
howpublished = {whataifound.org: A Registry of AI Scientific and Mathematical Discoveries},
note = {Result by Divergence Labs / Anthropic. Verification: Claimed. Autonomy: Search scaffold.},
url = {https://whataifound.org/finding/2026-10-27-ige-dislodging-antibody-designs}
}