Predicting Cell-Penetrating Peptide Uptake Mechanism from Sequence: A Machine Learning Approach

Predicting Cell-Penetrating Peptide Uptake Mechanism from Sequence: A Machine Learning Approach | 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 Method Article Predicting Cell-Penetrating Peptide Uptake Mechanism from Sequence: A Machine Learning Approach Nabil Brag This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8642513/v2 This work is licensed under a CC BY 4.0 License Status: Posted Version 2 posted You are reading this latest preprint version Show more versions Abstract Cell-penetrating peptides (CPPs) are promising drug delivery vectors, but their therapeutic efficacy depends critically on their cellular uptake mechanism. CPPs can enter cells via energy-dependent endocytosis or energy-independent direct translocation, with profound implications for cargo delivery and bioavailability. While numerous computational tools predict whether a peptide has cell-penetrating properties, none predict the uptake mechanism itself. Here, we present CPPMechPred, the first machine learning model specifically designed to predict CPP uptake mechanism from amino acid sequence. We curated a dataset of 142 CPPs with experimentally validated mechanisms from peer-reviewed literature. After removing sequences with >80% identity, 111 non-redundant peptides remained. Using nested 5-fold cross-validation with bootstrap confidence intervals, our best model achieved an AUC-ROC of 0.795 [95% CI: 0.711-0.872]. Feature importance analysis revealed that hydrophobicity, leucine content, and basic residue ratio are key predictors, consistent with known biophysical mechanisms. Error analysis identified hybrid peptides with intermediate properties as the main source of misclassification, reflecting the biological continuum between uptake mechanisms. Our model and dataset are freely available at https://github.com/Misterbra/cpp-mechanism-predictor . cell-penetrating peptides uptake mechanism endocytosis direct translocation machine learning peptide classification drug delivery SVM Figures Figure 1 Figure 2 Figure 3 Full Text Additional Declarations The authors declare no competing interests. Supplementary Files TableS1peptidereferences.csv Table S1. Complete list of 142 CPPs with sequences, uptake mechanisms, and literature references. Cite Share Download PDF Status: Posted Version 2 posted You are reading this latest preprint version Show more versions 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-8642513","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Method Article","associatedPublications":[],"authors":[{"id":577816320,"identity":"af78f5a4-b414-48e6-b361-97a4fbb0ff66","order_by":0,"name":"Nabil Brag","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABCElEQVRIiWNgGAWjYFACHoYDDxgOABnMByACzFAJ9gY8WhLAWtggFFwLzwHcWhggankMGFAU4dJizn72INCWO/n8/Ws+Pv7YxpC4nR2o92fOYQYeaex6LHvyEoBanlnOuPF2s8FBoJadzTwGjL3bgFr4ErBqMTiQYwDUctiA4cbZbRIgLRsOs6X/4AVqsefB7jCD828gWuRvnHn+A6olgfEvyBZcWm5AbTE438PGANHCfICZF6+Wd0C/GDwzMLzBZixx5pyEMViL7LZ0Hpxazuce/vCh4o6B3PnDDz9UlNnIbjh/sIHx7TZrOVxaoBqBWAIcPhJwMbwaIID/AGE1o2AUjIJRMDIBAEINZd/VFK+RAAAAAElFTkSuQmCC","orcid":"","institution":"","correspondingAuthor":true,"prefix":"","firstName":"Nabil","middleName":"","lastName":"Brag","suffix":""}],"badges":[],"createdAt":"2026-01-19 19:09:06","currentVersionCode":2,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-8642513/v2","doiUrl":"https://doi.org/10.21203/rs.3.rs-8642513/v2","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":103941095,"identity":"7d064cc2-9fb3-44f0-b35d-9c7939492eda","added_by":"auto","created_at":"2026-03-04 19:17:10","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":254646,"visible":true,"origin":"","legend":"\u003cp\u003eModel performance comparison with 95% confidence intervals from bootstrap resampling (n=1000).\u003c/p\u003e","description":"","filename":"erroranalysis.png","url":"https://assets-eu.researchsquare.com/files/rs-8642513/v2/ff3fe2a24e8b7e6e089fa7cc.png"},{"id":104401898,"identity":"aec81f47-6f97-4747-9c58-5feb60b80676","added_by":"auto","created_at":"2026-03-11 12:13:51","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":220442,"visible":true,"origin":"","legend":"\u003cp\u003et-SNE projection of CPPs based on sequence features, colored by mechanism. Overlap between classes reflects the biological continuum between uptake mechanisms.\u003c/p\u003e","description":"","filename":"tsnevisualization.png","url":"https://assets-eu.researchsquare.com/files/rs-8642513/v2/e3ae857d1a95ea6150e2c7b4.png"},{"id":103941093,"identity":"f367228d-bded-4fef-a9d3-108d84bed2b5","added_by":"auto","created_at":"2026-03-04 19:17:10","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":124343,"visible":true,"origin":"","legend":"\u003cp\u003eModel diagnostic plots from the non-redundant dataset (n=111). (A) ROC curve from aggregated 5-fold cross-validation predictions showing AUC of 0.795. (B) Confusion matrix showing sensitivity of 0.86 and specificity of 0.56.\u003c/p\u003e","description":"","filename":"rocconfusionmatrix.png","url":"https://assets-eu.researchsquare.com/files/rs-8642513/v2/69239dc29f7170db32e8c36d.png"},{"id":104408232,"identity":"ce98b911-5c49-4ffd-a585-9216bb2658f8","added_by":"auto","created_at":"2026-03-11 12:42:04","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1186034,"visible":true,"origin":"","legend":"","description":"","filename":"CPPMechPredResearchSquare.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8642513/v2_covered_d3fc12c8-0e49-40ee-bd62-f66ad64d4442.pdf"},{"id":103941092,"identity":"098f0f15-5107-4f5a-90be-e9c9a7ed0876","added_by":"auto","created_at":"2026-03-04 19:17:10","extension":"csv","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":9523,"visible":true,"origin":"","legend":"\u003cp\u003eTable S1. Complete list of 142 CPPs with sequences, uptake mechanisms, and literature references.\u003c/p\u003e","description":"","filename":"TableS1peptidereferences.csv","url":"https://assets-eu.researchsquare.com/files/rs-8642513/v2/81f5f6835faa940b1c9a0dbf.csv"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"Predicting Cell-Penetrating Peptide Uptake Mechanism from Sequence: A Machine Learning Approach","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","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":"cell-penetrating peptides, uptake mechanism, endocytosis, direct translocation, machine learning, peptide classification, drug delivery, SVM","lastPublishedDoi":"10.21203/rs.3.rs-8642513/v2","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8642513/v2","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eCell-penetrating peptides (CPPs) are promising drug delivery vectors, but their therapeutic efficacy depends critically on their cellular uptake mechanism. CPPs can enter cells via energy-dependent endocytosis or energy-independent direct translocation, with profound implications for cargo delivery and bioavailability. While numerous computational tools predict whether a peptide has cell-penetrating properties, none predict the uptake mechanism itself. Here, we present CPPMechPred, the first machine learning model specifically designed to predict CPP uptake mechanism from amino acid sequence. We curated a dataset of 142 CPPs with experimentally validated mechanisms from peer-reviewed literature. After removing sequences with \u0026gt;80% identity, 111 non-redundant peptides remained. Using nested 5-fold cross-validation with bootstrap confidence intervals, our best model achieved an AUC-ROC of 0.795 [95% CI: 0.711-0.872]. Feature importance analysis revealed that hydrophobicity, leucine content, and basic residue ratio are key predictors, consistent with known biophysical mechanisms. Error analysis identified hybrid peptides with intermediate properties as the main source of misclassification, reflecting the biological continuum between uptake mechanisms. Our model and dataset are freely available at \u003ca href=\"https://github.com/Misterbra/cpp-mechanism-predictor\" rel=\"noopener noreferrer\" target=\"_blank\"\u003ehttps://github.com/Misterbra/cpp-mechanism-predictor\u003c/a\u003e.\u003c/p\u003e","manuscriptTitle":"Predicting Cell-Penetrating Peptide Uptake Mechanism from Sequence: A Machine Learning Approach","msid":"","msnumber":"","nonDraftVersions":[{"code":2,"date":"2026-03-04 19:17:05","doi":"10.21203/rs.3.rs-8642513/v2","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}},{"code":1,"date":"2026-01-21 03:35:33","doi":"10.21203/rs.3.rs-8642513/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":"06db7c5a-9bad-483d-9dd9-c84ec08476e7","owner":[],"postedDate":"March 4th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-01-21T03:35:33+00:00","versionOfRecord":[],"versionCreatedAt":"2026-03-04 19:17:05","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v2","identity":"rs-8642513","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8642513","identity":"rs-8642513","version":["v2"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}