{"paper":{"title":"Towards A Generative Protein Evolution Machine with DPLM-Evo","license":"http://creativecommons.org/licenses/by/4.0/","headline":"DPLM-Evo models protein evolution by predicting explicit substitutions, insertions, and deletions in a discrete diffusion process.","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Jiasheng Ye, Liang Hong, Quanquan Gu, Shujian Huang, Xinyou Wang, Yu Li, Zaixiang Zheng","submitted_at":"2026-04-30T19:59:07Z","abstract_excerpt":"Proteins are shaped by gradual evolution under biophysical and functional constraints. Protein language models learn rich evolutionary constraints from large-scale sequences, and discrete diffusion-based protein language models~(\\eg, DPLMs) are promising for both understanding and generation. However, existing DPLMs typically rely on masked diffusion that contradicts a simple biological intuition: proteins evolve through accumulated edits, not by emerging from masks. Consequently, these frameworks lack explicit pretraining objectives for substitution and insertion/deletion (indel) operations, "},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"DPLM-Evo achieves state-of-the-art mutation effect prediction performance on ProteinGym in the single-sequence setting and enables variable-length simulated evolution and post-editing of existing proteins via explicit edit trajectories.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the contextualized evolutionary noising kernel produces biologically realistic, context-dependent mutation patterns and that decoupling the upsampled latent alignment space from the observed sequence space introduces no artifacts in indel generation.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"DPLM-Evo adds explicit edit operations and a latent alignment space to discrete diffusion protein models, achieving SOTA single-sequence mutation effect prediction on ProteinGym while supporting variable-length generation.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"DPLM-Evo models protein evolution by predicting explicit substitutions, insertions, and deletions in a discrete diffusion process.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"8f29473d4e5969e477019e1c6c6e1fb7cf2de2f134e99b84a5658bf73068daca"},"source":{"id":"2605.00182","kind":"arxiv","version":3},"verdict":{"id":"6bb3a6f9-6500-4b46-a372-4d3b30592439","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T21:00:16.456675Z","strongest_claim":"DPLM-Evo achieves state-of-the-art mutation effect prediction performance on ProteinGym in the single-sequence setting and enables variable-length simulated evolution and post-editing of existing proteins via explicit edit trajectories.","one_line_summary":"DPLM-Evo adds explicit edit operations and a latent alignment space to discrete diffusion protein models, achieving SOTA single-sequence mutation effect prediction on ProteinGym while supporting variable-length generation.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the contextualized evolutionary noising kernel produces biologically realistic, context-dependent mutation patterns and that decoupling the upsampled latent alignment space from the observed sequence space introduces no artifacts in indel generation.","pith_extraction_headline":"DPLM-Evo models protein evolution by predicting explicit substitutions, insertions, and deletions in a discrete diffusion process."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.00182/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-20T20:36:57.556032Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T18:25:12.708036Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"53a96496477894b02efaad47728fa3039158a268f66c60a970f7f79acf529ec2"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}