AnswersPublished14 sources
MMDesign Reports Binders in 18 of 19 De Novo Design Projects
MoleculeMind’s MMDesign reportedly found at least one binder in 18 of 19 projects (94.7%); 139 of 453 tested candidates bound (30.7%). The reported workflow narrows tens of thousands of computational designs to 9–50 candidates for lab testing, with a median of 21.
Published byEdited with GPT-6 LunaImages generated with GPT Image 2
Research answer
Key takeaways
- MoleculeMind’s MMDesign reportedly found at least one binder in 18 of 19 projects (94.7%); 139 of 453 tested candidates bound (30.7%).
- The reported workflow narrows tens of thousands of computational designs to 9–50 candidates for lab testing, with a median of 21.
- Coverage also reports binders for all six tested GPCR targets, with hit rates ranging from 12.1% to 100%.
AI Prompt
openai.comCreate a landscape editorial hero image for this Studio Global article: What does MoleculeMind’s October 29, 2026 technical report, available on Zenodo, reveal about MMDesign’s de novo nanobody (VHH) and macrocyc. Article summary: The reported results suggest MMDesign can repeatedly produce experimentally confirmed binders across different targets and two molecular formats, rather than merely generate plausible sequences. Available coverage corrob. Topic tags: general, general web, user generated, government, academic. Style: premium digital editorial illustration, source-backed research mood, clean composition, high detail, modern web publication hero. Use reference image context only for broad subject, composition, and topical grounding; do not copy the exact image. Avoid: logos, brand marks, copyrighted characters, real person likenesses, fake screenshots, UI text, readable text, wate
MoleculeMind’s MMDesign platform reportedly produced at least one lab-confirmed binder in 18 of 19 de novo design projects across 18 targets. That corresponds to a 94.7% project success rate. Separately, 139 of 453 tested candidates bound their targets, a 30.7% candidate hit rate—a different measure that shows most tested molecules did not bind. 17
18
The results, summarized in October 2026 coverage, span nanobodies (VHHs) and macrocyclic peptides. They point to a workflow designed to reduce how many candidates reach the lab, while leaving experimental testing as an essential part of discovery. 17
18
Two success rates answer different questions
A project counts as successful if it yields at least one confirmed binder. On that measure, MMDesign succeeded in 18 of 19 projects. The candidate hit rate instead counts the fraction of tested molecules that bound: 139 out of 453, or 30.7%. 17
18
Coverage also says 15 projects produced multiple binders. That matters because a project can produce more than one candidate for follow-up, even though the overall candidate hit rate is far below the project success rate. 18
How MMDesign narrows designs for testing
The reported process follows a generate–filter–validate sequence. MMDesign generates tens of thousands of computational candidates, applies filters for structural reliability, sequence naturalness and statistical-physics-based scoring, then sends 9–50 candidates per project for experimental testing; the reported median is 21. 17
18
The implication is not that computation replaces the lab. Rather, the approach aims to use computational screening to select a smaller experimental shortlist. The reported hit rate describes the tested candidates, not every sequence generated in silico. 17
18
Reported results include six GPCR targets
The October coverage says MMDesign produced binders for all six GPCR targets tested—CCR7, CCR8, EP2, GPR68, NK3R and A2AR—with reported candidate hit rates ranging from 12.1% to 100%. 17
19
GPCRs are membrane proteins, and their structures can present challenges for computational modeling and design. That makes the reported results across multiple GPCR targets notable as a design test, but they do not by themselves establish that the binders will work as medicines. 48
What the results do—and do not—establish
The reported project and candidate rates suggest MMDesign generated experimentally confirmed binders across multiple targets and two molecular formats. But the percentages are not interchangeable: 94.7% is the share of projects with at least one binder, while 30.7% is the share of tested candidates that bound. 17
18
The available October coverage provides headline results and workflow details, but not enough information here to assess every assay definition or compare hit rates directly across projects. The figures should therefore be read as reported outcomes, not as proof of therapeutic efficacy, safety or clinical readiness.
Studio Global AI
Continue your research
This page includes a source-backed answer you can continue inside Studio Global.
People also ask
What is the short answer to "MMDesign Reports Binders in 18 of 19 De Novo Design Projects"?
MoleculeMind’s MMDesign reportedly found at least one binder in 18 of 19 projects (94.7%); 139 of 453 tested candidates bound (30.7%).
What are the key points to validate first?
MoleculeMind’s MMDesign reportedly found at least one binder in 18 of 19 projects (94.7%); 139 of 453 tested candidates bound (30.7%). The reported workflow narrows tens of thousands of computational designs to 9–50 candidates for lab testing, with a median of 21.
What should I do next in practice?
Coverage also reports binders for all six tested GPCR targets, with hit rates ranging from 12.1% to 100%.