{"id":"https://openalex.org/W7222550199","doi":"https://doi.org/10.48550/arxiv.2610.21200","title":"Poisoning the corpora of protein design systems during pre-training and retrieval","display_name":"Poisoning the corpora of protein design systems during pre-training and retrieval","publication_year":2026,"publication_date":"2026-10-15","ids":{"openalex":"https://openalex.org/W7222550199","doi":"https://doi.org/10.48550/arxiv.2610.21200"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2610.21200","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.21200","pdf_url":"https://arxiv.org/pdf/2610.21200","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2610.21200","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5117779832","display_name":"Charalampos Koilakos","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I86519309","display_name":"The University of Texas at Austin","ror":"https://ror.org/00hj54h04","country_code":"US","type":"education","lineage":["https://openalex.org/I86519309"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Koilakos, Charalampos","raw_affiliation_strings":["Department of Computer Science , The University of Texas at Austin","Division of Pharmacology and Toxicology , College of Pharmacy , The University of Texas at Austin , Dell Pediatric Research Institute , Austin , TX , USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science , The University of Texas at Austin","institution_ids":["https://openalex.org/I86519309"]},{"raw_affiliation_string":"Division of Pharmacology and Toxicology , College of Pharmacy , The University of Texas at Austin , Dell Pediatric Research Institute , Austin , TX , USA","institution_ids":["https://openalex.org/I86519309"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5158861875","display_name":"Kimon-Antonios Provatas","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I86519309","display_name":"The University of Texas at Austin","ror":"https://ror.org/00hj54h04","country_code":"US","type":"education","lineage":["https://openalex.org/I86519309"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Provatas, Kimon-Antonios","raw_affiliation_strings":["Department of Computer Science , The University of Texas at Austin","Division of Pharmacology and Toxicology , College of Pharmacy , The University of Texas at Austin , Dell Pediatric Research Institute , Austin , TX , USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science , The University of Texas at Austin","institution_ids":["https://openalex.org/I86519309"]},{"raw_affiliation_string":"Division of Pharmacology and Toxicology , College of Pharmacy , The University of Texas at Austin , Dell Pediatric Research Institute , Austin , TX , USA","institution_ids":["https://openalex.org/I86519309"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5094583290","display_name":"Michail Patsakis","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I86519309","display_name":"The University of Texas at Austin","ror":"https://ror.org/00hj54h04","country_code":"US","type":"education","lineage":["https://openalex.org/I86519309"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Patsakis, Michail","raw_affiliation_strings":["Department of Computer Science , The University of Texas at Austin","Division of Pharmacology and Toxicology , College of Pharmacy , The University of Texas at Austin , Dell Pediatric Research Institute , Austin , TX , USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science , The University of Texas at Austin","institution_ids":["https://openalex.org/I86519309"]},{"raw_affiliation_string":"Division of Pharmacology and Toxicology , College of Pharmacy , The University of Texas at Austin , Dell Pediatric Research Institute , Austin , TX , USA","institution_ids":["https://openalex.org/I86519309"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5158873727","display_name":"Candace SY Chan","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I86519309","display_name":"The University of Texas at Austin","ror":"https://ror.org/00hj54h04","country_code":"US","type":"education","lineage":["https://openalex.org/I86519309"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Chan, Candace SY","raw_affiliation_strings":["Division of Pharmacology and Toxicology , College of Pharmacy , The University of Texas at Austin , Dell Pediatric Research Institute , Austin , TX , USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Division of Pharmacology and Toxicology , College of Pharmacy , The University of Texas at Austin , Dell Pediatric Research Institute , Austin , TX , USA","institution_ids":["https://openalex.org/I86519309"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5088991393","display_name":"Georgios A. Pavlopoulos","orcid":"https://orcid.org/0000-0002-4577-8276","observed_orcids":["https://orcid.org/0000-0002-4577-8276"]},"institutions":[{"id":"https://openalex.org/I4210113480","display_name":"Mohamed bin Zayed University of Artificial Intelligence","ror":"https://ror.org/0258gkt32","country_code":"AE","type":"education","lineage":["https://openalex.org/I4210113480"]},{"id":"https://openalex.org/I6820528","display_name":"Alexander Fleming Biomedical Sciences Research Center","ror":"https://ror.org/013x0ky90","country_code":"GR","type":"facility","lineage":["https://openalex.org/I6820528"]}],"countries":["AE","GR"],"is_corresponding":false,"raw_author_name":"Pavlopoulos, Georgios A.","raw_affiliation_strings":["Division of Biological and Life Sciences , Mohamed bin Zayed University of Artificial Intelligence (MBZUAI) , Building 1B , Masdar City , Abu Dhabi , United Arab Emirates","Institute for Fundamental Biomedical Research , BSRC \"Alexander Fleming , \" 16672 Vari , Greece"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Division of Biological and Life Sciences , Mohamed bin Zayed University of Artificial Intelligence (MBZUAI) , Building 1B , Masdar City , Abu Dhabi , United Arab Emirates","institution_ids":["https://openalex.org/I4210113480"]},{"raw_affiliation_string":"Institute for Fundamental Biomedical Research , BSRC \"Alexander Fleming , \" 16672 Vari , Greece","institution_ids":["https://openalex.org/I6820528"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5055883774","display_name":"Ilias Georgakopoulos\u2010Soares","orcid":"https://orcid.org/0000-0003-3641-1488","observed_orcids":["https://orcid.org/0000-0003-3641-1488"]},"institutions":[{"id":"https://openalex.org/I86519309","display_name":"The University of Texas at Austin","ror":"https://ror.org/00hj54h04","country_code":"US","type":"education","lineage":["https://openalex.org/I86519309"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Georgakopoulos-Soares, Ilias","raw_affiliation_strings":["Division of Pharmacology and Toxicology , College of Pharmacy , The University of Texas at Austin , Dell Pediatric Research Institute , Austin , TX , USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Division of Pharmacology and Toxicology , College of Pharmacy , The University of Texas at Austin , Dell Pediatric Research Institute , Austin , TX , USA","institution_ids":["https://openalex.org/I86519309"]}]}],"institutions":[],"countries_distinct_count":3,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.9994999766349792,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.9994999766349792,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T13851","display_name":"Law, AI, and Intellectual Property","score":9.999999747378752e-05,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11424","display_name":"Security and Verification in Computing","score":9.999999747378752e-05,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/data-poisoning-attacks","display_name":"data poisoning","score":0.7839999794960022},{"id":"https://openalex.org/keywords/protein-design","display_name":"protein design","score":0.7369999885559082},{"id":"https://openalex.org/keywords/protein-language-models","display_name":"protein language models","score":0.7440000176429749},{"id":"https://openalex.org/keywords/adversarial-triggers","display_name":"adversarial triggers","score":0.6389999985694885},{"id":"https://openalex.org/keywords/domain-transfer","display_name":"domain transfer","score":0.7210000157356262},{"id":"https://openalex.org/keywords/generative-ai-security","display_name":"generative AI security","score":0.5270000100135803}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5953999757766724},{"id":"https://openalex.org/C10010492","wikidata":"https://www.wikidata.org/wiki/Q3142557","display_name":"Protein sequencing","level":4,"score":0.526199996471405},{"id":"https://openalex.org/C36503486","wikidata":"https://www.wikidata.org/wiki/Q11235244","display_name":"Domain (mathematical analysis)","level":2,"score":0.5246000289916992},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.478300005197525},{"id":"https://openalex.org/C57273362","wikidata":"https://www.wikidata.org/wiki/Q576722","display_name":"Decoding methods","level":2,"score":0.46950000524520874},{"id":"https://openalex.org/C68387754","wikidata":"https://www.wikidata.org/wiki/Q7271585","display_name":"Schedule","level":2,"score":0.4431999921798706},{"id":"https://openalex.org/C39890363","wikidata":"https://www.wikidata.org/wiki/Q36108","display_name":"Generative grammar","level":2,"score":0.40149998664855957},{"id":"https://openalex.org/C144292202","wikidata":"https://www.wikidata.org/wiki/Q898273","display_name":"Protein domain","level":3,"score":0.3991999924182892},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3937000036239624},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.3578000068664551},{"id":"https://openalex.org/C37736160","wikidata":"https://www.wikidata.org/wiki/Q1801315","display_name":"Adversarial system","level":2,"score":0.3465000092983246},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.34459999203681946},{"id":"https://openalex.org/C2779786304","wikidata":"https://www.wikidata.org/wiki/Q5227243","display_name":"Data bank","level":2,"score":0.33629998564720154},{"id":"https://openalex.org/C152565575","wikidata":"https://www.wikidata.org/wiki/Q1124538","display_name":"Conditional random field","level":2,"score":0.3328999876976013},{"id":"https://openalex.org/C2985684807","wikidata":"https://www.wikidata.org/wiki/Q1513879","display_name":"Text generation","level":2,"score":0.2937000095844269},{"id":"https://openalex.org/C512654426","wikidata":"https://www.wikidata.org/wiki/Q19652","display_name":"Public domain","level":2,"score":0.28769999742507935},{"id":"https://openalex.org/C149364088","wikidata":"https://www.wikidata.org/wiki/Q185917","display_name":"Translation (biology)","level":4,"score":0.27639999985694885},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.2759999930858612},{"id":"https://openalex.org/C2988773926","wikidata":"https://www.wikidata.org/wiki/Q25104379","display_name":"Generative adversarial network","level":3,"score":0.274399995803833},{"id":"https://openalex.org/C147816474","wikidata":"https://www.wikidata.org/wiki/Q169525","display_name":"Protein engineering","level":3,"score":0.2606000006198883},{"id":"https://openalex.org/C40506919","wikidata":"https://www.wikidata.org/wiki/Q7452469","display_name":"Sequence learning","level":2,"score":0.25940001010894775}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2610.21200","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.21200","pdf_url":"https://arxiv.org/pdf/2610.21200","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2610.21200","doi":"https://doi.org/10.48550/arxiv.2610.21200","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2610.21200","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"public-domain","license_id":"https://openalex.org/licenses/public-domain","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2610.21200","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.21200","pdf_url":"https://arxiv.org/pdf/2610.21200","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[{"id":"https://openalex.org/G278569089","display_name":null,"funder_award_id":"23592","funder_id":"https://openalex.org/F4320327859","funder_display_name":"Hellenic Foundation for Research and Innovation"},{"id":"https://openalex.org/G2924758653","display_name":null,"funder_award_id":"23592-EMISSION","funder_id":"https://openalex.org/F4320327859","funder_display_name":"Hellenic Foundation for Research and Innovation"}],"funders":[{"id":"https://openalex.org/F4320327859","display_name":"Hellenic Foundation for Research and Innovation","ror":"https://ror.org/05v75r592"}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7222550199.pdf","grobid_xml":"https://content.openalex.org/works/W7222550199.grobid-xml"},"referenced_works_count":22,"referenced_works":["https://openalex.org/W1992950210","https://openalex.org/W2072058809","https://openalex.org/W2102461176","https://openalex.org/W2138122982","https://openalex.org/W2252830957","https://openalex.org/W2593864460","https://openalex.org/W2748789698","https://openalex.org/W2774423163","https://openalex.org/W2795038878","https://openalex.org/W3146944767","https://openalex.org/W3158635868","https://openalex.org/W4288066876","https://openalex.org/W4309506674","https://openalex.org/W4318071656","https://openalex.org/W4366352204","https://openalex.org/W4380989429","https://openalex.org/W4388024559","https://openalex.org/W4389519591","https://openalex.org/W4402264526","https://openalex.org/W4414581886","https://openalex.org/W7160282353","https://openalex.org/W7202018759"],"related_works":["https://openalex.org/W4388024559","https://openalex.org/W7164031929","https://openalex.org/W3015531900","https://openalex.org/W4283390570","https://openalex.org/W4312197262","https://openalex.org/W7135732278","https://openalex.org/W6887326600","https://openalex.org/W4409152738","https://openalex.org/W4383550741","https://openalex.org/W7162877092"],"abstract_inverted_index":{"Protein":[0],"foundation":[1],"models":[2,49],"increasingly":[3],"support":[4],"the":[5,24,42,51,60,107,119,130,181,234],"generation":[6,97,221],"of":[7,23,45,54,68,78,84,159,164,173,239],"novel":[8],"biological":[9,237],"sequences,":[10,177],"yet":[11],"their":[12],"reliance":[13],"on":[14,168],"large":[15],"public":[16],"datasets":[17],"creates":[18],"opportunities":[19],"for":[20,118,157,246],"adversarial":[21,207],"manipulation":[22],"material":[25],"they":[26],"learn":[27],"from":[28],"and":[29,50,65,100,162,199,236],"retrieve.":[30],"We":[31],"investigate":[32],"two":[33,79],"data-only":[34],"poisoning":[35,102],"attacks":[36],"against":[37],"protein":[38,47,57,74,208,220,240,248],"design":[39,249],"systems,":[40],"targeting":[41],"pre-training":[43,217],"corpus":[44,53,139],"autoregressive":[46],"language":[48],"retrieval":[52,138,155,197],"a":[55,137,153],"text-guided":[56],"generator.":[58],"For":[59,129],"first,":[61],"we":[62,133],"pre-trained":[63],"ProtGPT2":[64],"ProGen2":[66],"variants":[67],"300":[69],"million":[70],"parameters,":[71],"introducing":[72],"poisoned":[73,89,127,170],"sequences":[75],"containing":[76],"one":[77,83],"triggers":[80],"followed":[81],"by":[82,184],"three":[85],"payloads.":[86],"Across":[87],"all":[88],"runs,":[90],"both":[91],"architectures":[92],"acquired":[93,180],"highly":[94],"reliable":[95],"trigger-dependent":[96],"under":[98],"uniform":[99,120],"ramped":[101,105],"schedules.":[103],"Under":[104],"exposure,":[106],"earliest":[108,122],"majority":[109],"activation":[110,123],"occurred":[111,124],"after":[112],"only":[113],"200":[114],"contaminated":[115],"examples,":[116],"while":[117,178,224],"schedule":[121],"around":[125],"5000":[126],"samples.":[128],"second":[131],"attack,":[132],"inserted":[134,182],"records":[135,193],"into":[136],"system":[140],"without":[141],"modifying":[142],"its":[143],"weights":[144],"or":[145,218],"decoding":[146],"procedure.":[147],"A":[148,187],"single":[149],"query-aware":[150],"record":[151],"captured":[152],"top-16":[154],"position":[156],"58\\%":[158],"templated":[160],"requests":[161,174],"97\\%":[163],"free-text":[165],"requests.":[166],"Conditional":[167],"successful":[169],"retrieval,":[171],"62\\%":[172],"produced":[175],"altered":[176],"19-21%":[179],"domain":[183,200],"profile-HMM":[185],"homology.":[186],"universal":[188],"query-unaware":[189],"attack":[190],"using":[191],"1,500":[192],"also":[194],"achieved":[195],"substantial":[196],"capture":[198],"transfer.":[201],"Together,":[202],"these":[203],"findings":[204],"demonstrate":[205],"that":[206],"sequence":[209,241],"contributions":[210],"can":[211],"establish":[212],"persistent,":[213],"sequence-specific":[214],"behaviour":[215,228],"during":[216],"redirect":[219],"at":[222],"inference,":[223],"leaving":[225],"ordinary":[226],"model":[227],"largely":[229],"preserved.":[230],"Our":[231],"work":[232],"establishes":[233],"integrity":[235],"provenance":[238],"data":[242],"as":[243],"critical":[244],"safeguards":[245],"generative":[247],"systems.":[250]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-10-17T00:00:00"}