Multiomics integration of single-cell transcriptomics and bulk sequencing data identifies key biomarkers and...

Area of Science:

• Immunology and Molecular Biology
• Genomics and Bioinformatics
• Gastroenterology

Background:

• Inflammatory bowel disease (IBD), encompassing ulcerative colitis (UC) and Crohn's disease (CD), presents diagnostic challenges due to overlapping clinical presentations.
• Accurate molecular classification is crucial for advancing IBD subtype differentiation and creating reliable diagnostic methods.

Purpose of the Study:

• To identify distinct molecular patterns differentiating UC and CD subtypes using integrated peripheral blood single-cell RNA sequencing (scRNA-seq) and bulk sequencing data.
• To develop and validate machine learning-based diagnostic models for IBD subtypes.

Main Methods:

• Integration of scRNA-seq and bulk sequencing data from UC and CD patients.
• Application of machine learning algorithms and network analysis to identify hub genes and construct diagnostic models.
• Validation of identified biomarkers using an independent dataset and immunohistochemistry (IHC).

Main Results:

• Distinct immune signatures were observed: UC patients showed IL-17A+ effector memory T cells, while CD patients had IL-1β-producing immune cells.
• Cross-dataset integration identified 10 hub genes, and an 18-gene model achieved an AUC of 0.73 in validation.
• IHC confirmed subtype-specific protein expression for key genes like GZMA, PSME2, LDHB, and FOSB.

Conclusions:

• A multiomics approach successfully identified molecular signatures distinguishing UC from CD with moderate-to-good diagnostic capability (AUC=0.73).
• The findings highlight fundamental differences in immune regulation, metabolism, and tissue remodeling between IBD subtypes.
• The study provides a basis for developing blood-based diagnostic tools for IBD, with prospective clinical validation recommended.