Interpretable machine learning applied to high-dimensional salivary proteomics in pediatric inflammatory bowel diseases
Interpretierbares Machine Learning auf hochdimensionaler Speichelproteomik bei pädiatrischen entzündlichen Darmerkrankungen
Inflammatory Bowel Diseases · 09.11.2026
DOI: 10.1093/ibd/izag181 ↗Auf einen Blick
Eine Proof-of-Concept-Studie zeigte, dass sich pädiatrische IBD mit einer minimalen Protein-Signatur aus Speichel gut klassifizieren lassen. Das Modell erreichte 97,5 % Balanced Accuracy in Pilotproben.
Fragestellung
Lässt sich mit hochdimensionaler Speichelproteomik und interpretierbarem Machine Learning eine minimale Protein-Signatur zur Klassifikation pädiatrischer IBD definieren?
Methode
Proof-of-Concept-Studie mit unstimuliertem Speichel von pädiatrischen CD-, UC- und IBD-U-Patient:innen; Analyse mit NULISAseq 250 Protein Inflammation Panel, logistischer Regression, rekursiver Merkmalselimination und SHAP.
Ergebnisse
Differentiale Abundanzanalyse zwischen UC und CD ergab 53 signifikant unterschiedliche Proteine. Das ML-Modell identifizierte eine 14-Protein-Signatur und erreichte 97,5 % Balanced Accuracy in First-Visit-Pilotproben sowie 80,2 % mit Follow-up-Proben.
Schlussfolgerung der Autor:innen
Die Kombination aus Speichelproteomik und interpretierbarem AI/ML ermöglicht laut Autor:innen eine genaue, nicht-invasive Klassifikation pädiatrischer IBD mit minimalen Biomarker-Sets.
Kernaussagen
- 53 Proteine unterschieden sich zwischen UC und CD signifikant
- 14-Protein-Signatur für die Klassifikation identifiziert
- 97,5 % Balanced Accuracy in Pilotproben, 80,2 % mit Follow-up-Proben
Deutsche Zusammenfassung: KI-gestützt aus dem Original-Abstract erstellt; alle Zahlen wurden automatisch mit dem Original abgeglichen. Maßgeblich ist der Original-Abstract.
Kurz gefasst
Fachartikel in der Zeitschrift Inflammatory Bowel Diseases, online erschienen am 09.11.2026. Verfasst von Brittany T Rupp und 7 weiteren Autor:innen. Beteiligte Einrichtung: Department of Oral and Craniofacial Molecular Biology, Virginia Commonwealth University, Richmond, VA, United States. Die Arbeit nennt 56 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Crohn's and Colitis Foundation (Litwin IBD Pioneers Award, American Dental Association Science & Research Institute (startup funds & Volpe Research Scholar Award, Virginia Commonwealth University Department of Oral and Craniofacial Molecular Biology, Philips Institute for Oral Health Research.
Abstract (Originalsprache)
Background
Inflammatory bowel diseases (IBDs), including Crohn’s disease (CD), ulcerative colitis (UC), and inflammatory bowel disease unclassified (IBD-U), are chronic inflammatory disorders of the gastrointestinal tract. Current methods for classification and longitudinal monitoring are invasive, expensive, and often delayed, limiting timely diagnosis and management. This proof-of-concept study reports the first application of high-dimensional salivary proteomics integrated with interpretable artificial intelligence/machine learning (ML) to define a minimal protein signature for pediatric IBD classification with the ultimate goal of informing therapeutic decision-making.
Methods
Unstimulated saliva from pediatric CD, UC, and IBD-U patients was analyzed using Alamar Biosciences’ NULISAseq 250 Protein Inflammation Panel. Logistic regression with recursive feature elimination identified a candidate minimal discriminative signature. Performance was further assessed using follow-up samples. SHAP (SHapley Additive exPlanations) quantified patient-specific protein contributions and assessed biological similarity of IBD-U to CD and UC.
Results
Differential abundance analysis between UC and CD revealed 53 significantly different proteins. ML identified a 14-protein signature comprising chemokines/cytokines (CCL1, IFNA1-IFNA13, IL12p70, IL34, TNFSF11/RANKL), receptors/ligands (CD40LG, ICOSLG, IL1R2, IL17RA), structural/tissue-remodeling proteins (CD93, GFAP, SPP1), and growth factors/immune modulators (GDF2, GZMA). The proposed model achieved 97.5% balanced accuracy in first-visit pilot samples and 80.2% with matched follow-up samples. SHAP revealed patient-specific drivers and suggested biological alignment of IBD-U cases toward CD- or UC-like profiles.
Conclusions
This first-in-field integration of salivary proteomics with interpretable artificial intelligence/ML demonstrates that accurate, noninvasive classification of pediatric IBD is possible using minimal biomarker sets. This approach establishes a scalable framework for future longitudinal monitoring and supports earlier and more precise therapeutic interventions.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Oral and Craniofacial Molecular Biology, Virginia Commonwealth University, Richmond, VA, United States.
- Massey Comprehensive Cancer Center, Virginia Commonwealth University, Richmond, VA, United States.
- Division of Gastroenterology, Department of Pediatrics, University of North Carolina at Chapel Hill, Chapel Hill, NC, United States.
- Children's Research Institute, School of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, NC, United States.
- Department of Biostatistics, Virginia Commonwealth University, Richmond, VA, United States.
- Center for Gastrointestinal Biology and Disease, University of North Carolina at Chapel Hill, Chapel Hill, NC, United States.
ORCID-Profile: Brittany T Rupp, Kevin M Byrd
Förderung
- Crohn's and Colitis Foundation (Litwin IBD Pioneers Award
- American Dental Association Science & Research Institute (startup funds & Volpe Research Scholar Award
- Virginia Commonwealth University Department of Oral and Craniofacial Molecular Biology, Philips Institute for Oral Health Research
- Virginia Commonwealth University Wright Regional Center for Clinical & Translational Science Clinical and Translational Science
- National Cancer Institute/National Institutes of Health Cancer Center
- Crohn’s and Colitis Foundation (Litwin IBD Pioneers Award
Zitieren
APA
Rupp, B. T., Reyna, J., Giunta, A., Weaver, T., Chason, K., Liu, J., Gulati, A. S., & Byrd, K. M. (2026). Interpretable machine learning applied to high-dimensional salivary proteomics in pediatric inflammatory bowel diseases. Inflammatory Bowel Diseases. https://doi.org/10.1093/ibd/izag181
Vancouver
Rupp BT, Reyna J, Giunta A, Weaver T, Chason K, Liu J, et al. Interpretable machine learning applied to high-dimensional salivary proteomics in pediatric inflammatory bowel diseases. Inflammatory Bowel Diseases. 2026. doi:10.1093/ibd/izag181
BibTeX
@article{Rupp2026,
author = {Rupp, Brittany T and Reyna, Joaquin and Giunta, Ally and Weaver, Theresa and Chason, Kelly and Liu, Jinze and Gulati, Ajay S and Byrd, Kevin M},
title = {Interpretable machine learning applied to high-dimensional salivary proteomics in pediatric inflammatory bowel diseases},
journal = {Inflammatory Bowel Diseases},
year = {2026},
doi = {10.1093/ibd/izag181}
}
EUVIMED