{"id":"https://openalex.org/W7225622293","doi":"https://doi.org/10.21203/rs.3.rs-11270647/v1","title":"Negative Extrapolation for Targeted Motif Suppression in Molecular Generators","display_name":"Negative Extrapolation for Targeted Motif Suppression in Molecular Generators","publication_year":2026,"publication_date":"2026-11-09","ids":{"openalex":"https://openalex.org/W7225622293","doi":"https://doi.org/10.21203/rs.3.rs-11270647/v1"},"language":"en","primary_location":{"id":"doi:10.21203/rs.3.rs-11270647/v1","doi":"https://doi.org/10.21203/rs.3.rs-11270647/v1","is_oa":true,"landing_page_url":"https://doi.org/10.21203/rs.3.rs-11270647/v1","pdf_url":"https://www.researchsquare.com/article/rs-11270647/latest.pdf","source":{"id":"https://openalex.org/S4306525896","display_name":"Research Square","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"Research Square","raw_type":"posted-content"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://www.researchsquare.com/article/rs-11270647/latest.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5088042767","display_name":"J. Hesse","orcid":"https://orcid.org/0009-0004-1819-7726","observed_orcids":["https://orcid.org/0009-0004-1819-7726"]},"institutions":[{"id":"https://openalex.org/I62916508","display_name":"Technical University of Munich","ror":"https://ror.org/02kkvpp62","country_code":"DE","type":"education","lineage":["https://openalex.org/I62916508"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Joshua Hesse","raw_affiliation_strings":["Technical University of Munich"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Technical University of Munich","institution_ids":["https://openalex.org/I62916508"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5120369962","display_name":"Julia Westermayr","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4210156779","display_name":"Leibniz Institute of Surface Engineering","ror":"https://ror.org/04vx4mk32","country_code":"DE","type":"facility","lineage":["https://openalex.org/I315704651","https://openalex.org/I4210156779"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Julia Westermayr","raw_affiliation_strings":["Leibniz Institute of Surface Engineering"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Leibniz Institute of Surface Engineering","institution_ids":["https://openalex.org/I4210156779"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5048928672","display_name":"Stephan A. Sieber","orcid":"https://orcid.org/0000-0002-9400-906X","observed_orcids":["https://orcid.org/0000-0002-9400-906X"]},"institutions":[{"id":"https://openalex.org/I62916508","display_name":"Technical University of Munich","ror":"https://ror.org/02kkvpp62","country_code":"DE","type":"education","lineage":["https://openalex.org/I62916508"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Stephan A. Sieber","raw_affiliation_strings":["Technical University of Munich"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Technical University of Munich","institution_ids":["https://openalex.org/I62916508"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"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/T10211","display_name":"Computational Drug Discovery Methods","score":0.9970999956130981,"subfield":{"id":"https://openalex.org/subfields/3002","display_name":"Drug Discovery"},"field":{"id":"https://openalex.org/fields/30","display_name":"Pharmacology, Toxicology and Pharmaceutics"},"domain":{"id":"https://openalex.org/domains/1","display_name":"Life Sciences"}},"topics":[{"id":"https://openalex.org/T10211","display_name":"Computational Drug Discovery Methods","score":0.9970999956130981,"subfield":{"id":"https://openalex.org/subfields/3002","display_name":"Drug Discovery"},"field":{"id":"https://openalex.org/fields/30","display_name":"Pharmacology, Toxicology and Pharmaceutics"},"domain":{"id":"https://openalex.org/domains/1","display_name":"Life Sciences"}},{"id":"https://openalex.org/T11948","display_name":"Machine Learning in Materials Science","score":0.000699999975040555,"subfield":{"id":"https://openalex.org/subfields/2505","display_name":"Materials Chemistry"},"field":{"id":"https://openalex.org/fields/25","display_name":"Materials Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.00019999999494757503,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/flow-matching","display_name":"flow matching","score":0.43799999356269836},{"id":"https://openalex.org/keywords/molecular-diversity","display_name":"molecular diversity","score":0.3409999907016754},{"id":"https://openalex.org/keywords/chemical-space","display_name":"chemical space","score":0.41100001335144043}],"concepts":[{"id":"https://openalex.org/C99726746","wikidata":"https://www.wikidata.org/wiki/Q906396","display_name":"Chemical space","level":3,"score":0.6934000253677368},{"id":"https://openalex.org/C132459708","wikidata":"https://www.wikidata.org/wiki/Q744069","display_name":"Extrapolation","level":2,"score":0.6000999808311462},{"id":"https://openalex.org/C32276052","wikidata":"https://www.wikidata.org/wiki/Q908349","display_name":"Motif (music)","level":2,"score":0.5019999742507935},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.49970000982284546},{"id":"https://openalex.org/C186060115","wikidata":"https://www.wikidata.org/wiki/Q30336093","display_name":"Biological system","level":1,"score":0.4819999933242798},{"id":"https://openalex.org/C2780992000","wikidata":"https://www.wikidata.org/wiki/Q17016113","display_name":"Generator (circuit theory)","level":3,"score":0.43220001459121704},{"id":"https://openalex.org/C132677234","wikidata":"https://www.wikidata.org/wiki/Q3273544","display_name":"Structural motif","level":2,"score":0.4235999882221222},{"id":"https://openalex.org/C32909587","wikidata":"https://www.wikidata.org/wiki/Q11369","display_name":"Molecule","level":2,"score":0.42250001430511475},{"id":"https://openalex.org/C185592680","wikidata":"https://www.wikidata.org/wiki/Q2329","display_name":"Chemistry","level":0,"score":0.41609999537467957},{"id":"https://openalex.org/C161624437","wikidata":"https://www.wikidata.org/wiki/Q1988322","display_name":"Small molecule","level":2,"score":0.3905999958515167},{"id":"https://openalex.org/C167966045","wikidata":"https://www.wikidata.org/wiki/Q5532625","display_name":"Generative model","level":3,"score":0.3887999951839447},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.35280001163482666},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3140999972820282},{"id":"https://openalex.org/C70721500","wikidata":"https://www.wikidata.org/wiki/Q177005","display_name":"Computational biology","level":1,"score":0.3102000057697296},{"id":"https://openalex.org/C2775905019","wikidata":"https://www.wikidata.org/wiki/Q192572","display_name":"In silico","level":3,"score":0.3082999885082245},{"id":"https://openalex.org/C2776436953","wikidata":"https://www.wikidata.org/wiki/Q5163215","display_name":"Consistency (knowledge bases)","level":2,"score":0.3066999912261963},{"id":"https://openalex.org/C21951064","wikidata":"https://www.wikidata.org/wiki/Q899212","display_name":"Combinatorial chemistry","level":1,"score":0.29499998688697815},{"id":"https://openalex.org/C134121241","wikidata":"https://www.wikidata.org/wiki/Q899301","display_name":"Yield (engineering)","level":2,"score":0.28940001130104065},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.25940001010894775}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.21203/rs.3.rs-11270647/v1","doi":"https://doi.org/10.21203/rs.3.rs-11270647/v1","is_oa":true,"landing_page_url":"https://doi.org/10.21203/rs.3.rs-11270647/v1","pdf_url":"https://www.researchsquare.com/article/rs-11270647/latest.pdf","source":{"id":"https://openalex.org/S4306525896","display_name":"Research Square","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"Research Square","raw_type":"posted-content"}],"best_oa_location":{"id":"doi:10.21203/rs.3.rs-11270647/v1","doi":"https://doi.org/10.21203/rs.3.rs-11270647/v1","is_oa":true,"landing_page_url":"https://doi.org/10.21203/rs.3.rs-11270647/v1","pdf_url":"https://www.researchsquare.com/article/rs-11270647/latest.pdf","source":{"id":"https://openalex.org/S4306525896","display_name":"Research Square","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"Research Square","raw_type":"posted-content"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320311647","display_name":"Merck KGaA","ror":"https://ror.org/04b2dty93"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7225622293.pdf","grobid_xml":"https://content.openalex.org/works/W7225622293.grobid-xml"},"referenced_works_count":39,"referenced_works":["https://openalex.org/W2033495141","https://openalex.org/W2034549041","https://openalex.org/W2043509228","https://openalex.org/W2060531713","https://openalex.org/W2066273100","https://openalex.org/W2069126953","https://openalex.org/W2332896222","https://openalex.org/W2529996553","https://openalex.org/W2578240541","https://openalex.org/W2610148085","https://openalex.org/W2626778328","https://openalex.org/W2773987374","https://openalex.org/W2774836366","https://openalex.org/W2887447356","https://openalex.org/W2900090807","https://openalex.org/W2900694120","https://openalex.org/W2908510526","https://openalex.org/W2909063104","https://openalex.org/W2945551948","https://openalex.org/W2949986955","https://openalex.org/W2950635152","https://openalex.org/W2970971581","https://openalex.org/W3034806393","https://openalex.org/W3165171933","https://openalex.org/W4310998073","https://openalex.org/W4319310661","https://openalex.org/W4389696354","https://openalex.org/W4390871724","https://openalex.org/W4390874260","https://openalex.org/W4392015344","https://openalex.org/W4416052399","https://openalex.org/W4416204720","https://openalex.org/W4416372976","https://openalex.org/W6929394108","https://openalex.org/W7119846043","https://openalex.org/W7133183305","https://openalex.org/W7133216281","https://openalex.org/W7161991326","https://openalex.org/W7211903761"],"related_works":["https://openalex.org/W4214884080","https://openalex.org/W4387051186","https://openalex.org/W7223130300","https://openalex.org/W4283032374","https://openalex.org/W4226198901","https://openalex.org/W7083624016","https://openalex.org/W4313236609","https://openalex.org/W2808177712","https://openalex.org/W4391641683","https://openalex.org/W4414631713"],"abstract_inverted_index":{"Abstract":[0],"Broadly":[1],"trained":[2],"molecular":[3,228,254,266,291],"generators":[4,255],"can":[5,323],"cover":[6],"diverse":[7],"chemical":[8,60,233,300],"space,":[9],"but":[10],"may":[11],"also":[12],"produce":[13],"motifs":[14,95,106,141,211],"that":[15,225],"are":[16],"undesirable":[17,57],"in":[18,84,290],"a":[19,27,50,80,86,120,126,248,271,283,295],"given":[20],"discovery":[21],"context,":[22],"e.g.,":[23],"reactive":[24],"motifs.":[25],"Adapting":[26],"pretrained":[28,122],"generator":[29,87],"to":[30,103,186,214,240,286,337],"suppress":[31,104],"such":[32],"chemistry":[33],"is":[34,88,101,344],"therefore":[35],"an":[36,56,287],"important":[37,288],"practical":[38],"problem.":[39],"However,":[40],"most":[41],"model":[42,191,339],"steering":[43],"methods":[44],"enrich":[45],"desirable":[46],"compounds":[47],"or":[48],"optimize":[49],"predefined":[51],"reward,":[52],"whereas":[53],"selectively":[54,252,293],"suppressing":[55,294],"region":[58,298],"of":[59,68,153,179,203,208,221,281,299,308,332],"space":[61,301],"without":[62,138,269],"compromising":[63],"the":[64,69,97,109,139,146,151,194,201,209,218,278,304,309],"quality":[65,305],"and":[66,96,125,133,161,177,216,230,234,258,261,268,306,314,330],"diversity":[67,267,307],"remaining":[70,310],"distribution":[71],"remains":[72],"challenging.":[73],"Here,":[74],"we":[75,317],"adapt":[76],"negative":[77],"extrapolation":[78,343],"(NE),":[79],"parameter-space":[81],"editing":[82,176,253,329],"strategy":[83],"which":[85],"fine-tuned":[89],"on":[90,136],"molecules":[91,137,205,224],"containing":[92,206],"unwanted":[93,140,210,297],"structural":[94,155],"resulting":[98],"weight":[99],"update":[100,181,285,335],"reversed,":[102],"these":[105,243],"directly":[107],"within":[108],"generator.":[110],"We":[111,169,276],"evaluated":[112],"NE":[113,149,172,199,246],"using":[114],"two":[115],"custom":[116],"string-based":[117,147,260,313],"autoregressive":[118],"generators,":[119,148,316],"public":[121],"recurrent":[123],"prior,":[124],"three-dimensional":[127,195,232,262,315],"flow-matching":[128,196],"generator,":[129,197],"with":[130],"unedited":[131],"generation":[132],"conventional":[134,167],"fine-tuning":[135],"as":[142,247],"principal":[143],"comparators.":[144],"Across":[145,312],"reduced":[150,200],"occurrence":[152],"four":[154],"motif":[156],"classes":[157],"(reactive,":[158],"metal-binding,":[159],"charged,":[160],"assay-interference":[162],"motifs)":[163],"more":[164],"strongly":[165],"than":[166],"fine-tuning.":[168],"further":[170,318],"extended":[171],"through":[173,326],"selective":[174,327],"parameter":[175,321,328],"combinations":[178,331],"learned":[180,284,334],"directions,":[182],"enabling":[183],"generative":[184],"validity":[185,340],"be":[187,324],"preserved":[188],"across":[189,256,259],"different":[190],"architectures.":[192],"In":[193],"corrected":[198],"fraction":[202],"valid":[204,227],"any":[207],"from":[212,238],"31.1%":[213],"2.2%":[215],"increased":[217],"fixed-budget":[219],"yield":[220],"unique,":[222],"motif-free":[223],"were":[226],"graphs":[229],"passed":[231],"geometric":[235],"consistency":[236],"checks":[237],"60.0%":[239],"80.8%.":[241],"Together,":[242],"results":[244],"support":[245],"simple":[249],"approach":[250],"for":[251],"architectures":[257],"representations,":[263],"while":[264,302],"preserving":[265,303],"requiring":[270],"reward-optimization":[272],"loop.":[273],"Scientific":[274],"Contribution:":[275],"apply":[277],"parameter-arithmetic":[279],"principle":[280],"reversing":[282],"problem":[289],"generation:":[292],"defined":[296],"distribution.":[311],"investigate":[319],"how":[320],"arithmetic":[322],"adapted":[325],"independently":[333],"directions":[336],"retain":[338],"when":[341],"full-model":[342],"disruptive.":[345]},"counts_by_year":[],"updated_date":"2026-11-11T12:16:26.678958","created_date":"2026-11-10T00:00:00"}