LPT: bounded agentic LLM loops for de novo protein binder design
The LPT preprint describes fixed, bounded LLM-agent stages with access to a literature database of about 14,000 papers. A database-backed model answered 41 of 48 benchmark questions, versus none without access. Three binder-design demonstrations targeted CALCRL/RAMP1, PD-1/PD-L1 and PVR–TIGIT, with evaluation by other folding models. Archived as a computational workflow/method benchmark: the reviewed abstract establishes no binding measurement or separately validated scientific result.
Technology: LPT framework: fixed orchestration of bounded LLM agentic loops with literature database and design tools
Publication: Preprint · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
Result, AI contribution and limits
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
- Publication status: Preprint
- Test type: retrospective or benchmark
- Tested: Question-answering benchmark and three computational binder-design campaigns checked using multiple structure-prediction models.
- Still unknown: Abstract and original-server metadata reviewed. No claim that the inaccessible full paper contains no experiments is made; the reviewed abstract reports only in-silico performance.
- AI role: LLM stages identified the design intent, selected targets, structures and hotspot residues, and designed candidate binders.
Evidence sources: https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=DOI%3A%2210.64898%2F2026.10.29.762730%22&resultType=core&format=json · https://api.biorxiv.org/details/biorxiv/10.64898/2026.10.29.762730
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review scope and date are not recorded; source checked does not establish full-text review.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.
Follow-up and relationships
No follow-up recorded; this is not a claim that no follow-up exists.
Follow-up review scope and date are not recorded.
No source-supported relationships recorded.