IBD Subtype Biomarkers: Multiomics Integration of Single-Cell and Bulk Sequencing Data

43 consistently dysregulated genes and 10 hub genes were identified by integrating peripheral blood single-cell RNA sequencing and bulk sequencing data from UC and CD patients. Linghui Song published this analysis on figshare in March 2026. An 18-gene machine learning model achieved an AUC of 0.73 in independent validation.

Use Cases

• Train subtype classification models based on identified hub genes like THBS1 and PLAUR.
• Validate biomarker expression using immunohistochemical confirmation data referenced in the description.
• Analyze immune cell signatures distinguishing UC and CD based on IL-17A+ effector memory T cells and IL-1β-producing cells.
• Investigate cross-dataset integration methods for identifying consistently dysregulated genes.
• Evaluate diagnostic model performance using the reported AUC score from independent validation.

Strengths

• Identified 43 consistently dysregulated genes and 10 hub genes.
• Model validation achieved an AUC of 0.73 using an independent dataset.
• Biomarker expression was confirmed via immunohistochemistry in clinical samples.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The dataset is 286.0 KB, indicating a limited scope of raw data.

Provenance

Source

figshare

Collection Method

Integration of public scRNA-seq (GSE125527) and bulk sequencing (GSE75214) datasets, with validation using GSE179285.

Freshness

Last updated 2026-03-19 06:26:11; freshness should be verified.