{"paper":{"title":"ProtDBench: A Unified Benchmark of Protein Binder Design and Evaluation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"A new benchmark for protein binder design shows that common structure prediction models disagree substantially on which designs succeed under identical rules.","cross_cats":["cs.AI"],"primary_cat":"q-bio.QM","authors_text":"Chengyue Gong, Cong Liu, Jiaqi Guan, Jinyuan Sun, Milong Ren, Wenzhi Xiao, Xinshi Chen","submitted_at":"2026-05-05T11:48:59Z","abstract_excerpt":"Recent advances in de novo protein binder design have enabled increasing experimental validation, yet reported in silico metrics remain difficult to interpret or compare across studies due to non-standardized evaluation protocols. We introduce ProtDBench, a standardized and throughput-aware evaluation framework for protein binder design. ProtDBench defines unified benchmark tasks, evaluation protocols, and success criteria, enabling systematic analysis of how evaluation design influences observed performance. Using a large wet-lab annotated dataset, we analyze commonly used structure predictio"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Using a large wet-lab annotated dataset, we analyze commonly used structure prediction models as evaluation verifiers, revealing substantial verifier-dependent bias and limited agreement under identical filtering protocols. ... Overall, ProtDBench provides a fair and reproducible evaluation pipeline that supports systematic and controlled comparison of protein binder design methods under realistic evaluation settings.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the defined success criteria, filtering protocols, and wet-lab annotated dataset accurately reflect real experimental binder performance and provide a representative basis for benchmarking.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"ProtDBench standardizes protein binder design evaluation using wet-lab data, exposing verifier biases, metric dependencies, and trade-offs between success rate, speed, and structural diversity.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A new benchmark for protein binder design shows that common structure prediction models disagree substantially on which designs succeed under identical rules.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"75e6ef31d6837543ba092d650fe111d71d724b28c36cea7f8a9cd14f537e5132"},"source":{"id":"2605.04118","kind":"arxiv","version":2},"verdict":{"id":"f144bf74-40b8-4ca1-88b0-b1ae56f84e27","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-08T18:09:03.659188Z","strongest_claim":"Using a large wet-lab annotated dataset, we analyze commonly used structure prediction models as evaluation verifiers, revealing substantial verifier-dependent bias and limited agreement under identical filtering protocols. ... Overall, ProtDBench provides a fair and reproducible evaluation pipeline that supports systematic and controlled comparison of protein binder design methods under realistic evaluation settings.","one_line_summary":"ProtDBench standardizes protein binder design evaluation using wet-lab data, exposing verifier biases, metric dependencies, and trade-offs between success rate, speed, and structural diversity.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the defined success criteria, filtering protocols, and wet-lab annotated dataset accurately reflect real experimental binder performance and provide a representative basis for benchmarking.","pith_extraction_headline":"A new benchmark for protein binder design shows that common structure prediction models disagree substantially on which designs succeed under identical rules."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.04118/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-20T13:36:39.248387Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_title_agreement","ran_at":"2026-05-20T00:31:21.665980Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T15:07:27.199343Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"9522f94c1f930f5201dea4bd8886f4758b5b7e2598cc6a2a986b8e861c421d81"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":3,"snapshot_sha256":"ff35b7430ce1419199860890142a12cc8dd08fe311b31a58b6e9832801aa4498"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}