ARIA: Adaptive Reasoning for Integrated Analysis — An LLM-Powered Framework for Autonomous Transcriptome Analysis with Decision-Aware Workflow Orchestration

ARIA: Adaptive Reasoning for Integrated Analysis — An LLM-Powered Framework for Autonomous Transcriptome Analysis with Decision-Aware Workflow Orchestration | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article ARIA: Adaptive Reasoning for Integrated Analysis — An LLM-Powered Framework for Autonomous Transcriptome Analysis with Decision-Aware Workflow Orchestration Byeongsoo Kang This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9500973/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract RNA-seq transcriptome analysis requires a multi-step workflow involving quality control, alignment, quantification, differential expression testing, pathway analysis, and biological interpretation. While automated pipelines such as nf-core/rnaseq execute these steps reproducibly, the critical decisions between steps — evaluating quality metrics, selecting statistical methods, adapting analysis strategies based on intermediate results, and interpreting findings in biological context — remain dependent on expert bioinformaticians. Here we present ARIA (Adaptive Reasoning for Integrated Analysis), an open-source framework that uses a Large Language Model (LLM) as a reasoning engine to autonomously navigate the decision space of RNA-seq analysis. ARIA implements eight Decision Points (DPs) that govern quality assessment, strategy adaptation, method selection, and result interpretation, combining rule-based thresholds with LLM-driven contextual reasoning. We benchmark ARIA on four public RNA-seq datasets spanning three species: SEQC (GSE49712, human, 10,430 DEGs), Airway (GSE52778, human paired design, 951 DEGs), Fmr1 KO (GSE180135, mouse, 398 DEGs), and Pasilla (GSE18508, Drosophila, 224 DEGs with mixed library types). ARIA correctly identifies paired experimental designs (increasing DEG detection by 21–47%), detects technical covariates such as library type (+4–30% DEG gain), adaptively selects analysis strategies based on DEG counts, and cross-validates results across DESeq2/edgeR/limma-voom (r > 0.99). All 7/7 known dexamethasone-responsive genes (Himes et al., 2014) were recovered in Airway, 9/11 FMRP translational targets (Bhatt et al., 2012; Darnell et al., 2011) in Fmr1 KO, and 15/16 tissue-type markers (9/9 brain-enriched genes + 6/7 cancerassociated genes; SEQC Consortium, 2014) in SEQC were correctly assigned (BCL2 showed brainenriched expression despite being classified as a cancer marker). ARIA is freely available at https://github.com/shoo99/ARIA. Bioinformatics Computational Biology Artificial Intelligence and Machine Learning Epigenetics & Genomics RNA-seq differential expression large language models autonomous analysis decision-aware pipeline transcriptomics Full Text Additional Declarations The authors declare no competing interests. Supplementary Files ARIAsupplementary1.pdf Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9500973","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":628050362,"identity":"3913185a-c139-4ffe-8329-0b60af8726d1","order_by":0,"name":"Byeongsoo Kang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAwElEQVRIiWNgGAWjYDACCSD+UGEDJBkbDxCthXHGmTSQlgbitTDzth0Gs4nTIj+7+eBj3rbzdmvbDwNtqbGJJqjF4M6xZMM5524nbzuTCNRyLC23gaAWiRwziTdlt5PNDgC1MDYcJqxFfgZQCw/buWSz8w+J1MJwI8dMkqftgJ3ZDWJtMbiRlmw440xygtkNoC0JxPhFfkbywQcfKuzszc6nP3zwocaGCIdBQSJYZQKxykHAnhTFo2AUjIJRMMIAANHaSZTsx3+MAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0009-0007-2324-2351","institution":"SYSOFT","correspondingAuthor":true,"prefix":"","firstName":"Byeongsoo","middleName":"","lastName":"Kang","suffix":""}],"badges":[],"createdAt":"2026-04-23 01:13:22","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-9500973/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9500973/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":107870759,"identity":"36893b14-44a9-4327-9a7f-3781751eae4c","added_by":"auto","created_at":"2026-04-27 07:40:33","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1316353,"visible":true,"origin":"","legend":"","description":"","filename":"ARIAbioRxiv8.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9500973/v1_covered_b6ea54d8-eed1-4c60-8ea7-c3164366720f.pdf"},{"id":107861098,"identity":"7e609787-423f-439e-ac25-6b9d0c9eb75d","added_by":"auto","created_at":"2026-04-27 05:33:27","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":28563,"visible":true,"origin":"","legend":"","description":"","filename":"ARIAsupplementary1.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9500973/v1/c9902ca87c3ee059bc919a64.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003eARIA: Adaptive Reasoning for Integrated Analysis — An LLM-Powered Framework for Autonomous Transcriptome Analysis with Decision-Aware Workflow Orchestration\u003c/p\u003e","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"SYSOFT","isAcceptedByJournal":false,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"RNA-seq, differential expression, large language models, autonomous analysis, decision-aware pipeline, transcriptomics","lastPublishedDoi":"10.21203/rs.3.rs-9500973/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9500973/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eRNA-seq transcriptome analysis requires a multi-step workflow involving quality control, alignment, quantification, differential expression testing, pathway analysis, and biological interpretation. While automated pipelines such as nf-core/rnaseq execute these steps reproducibly, the critical decisions between steps — evaluating quality metrics, selecting statistical methods, adapting analysis strategies based on intermediate results, and interpreting findings in biological context — remain dependent on expert bioinformaticians. Here we present ARIA (Adaptive Reasoning for Integrated Analysis), an open-source framework that uses a Large Language Model (LLM) as a reasoning engine to autonomously navigate the decision space of RNA-seq analysis. ARIA implements eight Decision Points (DPs) that govern quality assessment, strategy adaptation, method selection, and result interpretation, combining rule-based thresholds with LLM-driven contextual reasoning. We benchmark ARIA on four public RNA-seq datasets spanning three species: SEQC (GSE49712, human, 10,430 DEGs), Airway (GSE52778, human paired design, 951 DEGs), Fmr1 KO (GSE180135, mouse, 398 DEGs), and Pasilla (GSE18508, Drosophila, 224 DEGs with mixed library types). ARIA correctly identifies paired experimental designs (increasing DEG detection by 21–47%), detects technical covariates such as library type (+4–30% DEG gain), adaptively selects analysis strategies based on DEG counts, and cross-validates results across DESeq2/edgeR/limma-voom (r \u0026gt; 0.99). All 7/7 known dexamethasone-responsive genes (Himes et al., 2014) were recovered in Airway, 9/11 FMRP translational targets (Bhatt et al., 2012; Darnell et al., 2011) in Fmr1 KO, and 15/16 tissue-type markers (9/9 brain-enriched genes + 6/7 cancerassociated genes; SEQC Consortium, 2014) in SEQC were correctly assigned (BCL2 showed brainenriched expression despite being classified as a cancer marker). ARIA is freely available at https://github.com/shoo99/ARIA.\u003c/p\u003e","manuscriptTitle":"ARIA: Adaptive Reasoning for Integrated Analysis — An LLM-Powered Framework for Autonomous Transcriptome Analysis with Decision-Aware Workflow Orchestration","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-27 05:33:23","doi":"10.21203/rs.3.rs-9500973/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"492e6774-0079-4904-981b-ea51c616051c","owner":[],"postedDate":"April 27th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":67049870,"name":"Bioinformatics"},{"id":67049871,"name":"Computational Biology"},{"id":67049872,"name":"Artificial Intelligence and Machine Learning"},{"id":67049873,"name":"Epigenetics \u0026 Genomics"}],"tags":[],"updatedAt":"2026-04-27T05:33:23+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-27 05:33:23","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9500973","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9500973","identity":"rs-9500973","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}