Oral Submucosal Fibrosis Multi-Omics Data for Biomarker Discovery via Machine Learning

Chinmay Nitin Mokal published this dataset on figshare in May 2026. It contains results from an integrative multi-omics analysis of Oral Submucosal Fibrosis (OSF), including RNA-seq, microarray, epigenomic, and single-cell RNA-seq data. The study identified differentially expressed genes and used machine learning to find potential diagnostic biomarkers and therapeutic targets.

Use Cases

• Training machine learning models for disease classification based on the top 5 gene features identified.
• Validating potential diagnostic biomarkers using the 11 gene signatures from the integrative transcriptome and methylome analysis.
• Analyzing immune cell infiltration patterns in OSF based on the CIBERSORTx enrichment results for T cells, mast cells, and macrophages.
• Investigating cell-type-specific gene expression using the single-cell RNA-seq analysis results for keratinocytes, epithelial cells, and dendritic cells.

Strengths

• Integrates 7 distinct datasets (3 RNA-seq, 2 microarray, 1 epigenomic, 1 single-cell RNA-seq) from the Gene Expression Omnibus.
• Machine learning validation reported high performance, with a Random Forest model achieving an AUROC and AUPRC of 0.99.
• Analysis identified 44 differentially expressed genes (29 upregulated, 15 downregulated) and 11 key biomarker genes.
• All analysis codes and machine learning models are provided in a linked GitHub repository.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data may reflect bias inherent to the specific datasets sourced from the Gene Expression Omnibus.

Provenance

Source

Gene Expression Omnibus (GEO) database

Collection Method

Computational analysis of downloaded public datasets using DESeq2, CIBERSORTx, and machine learning models.

Time Range

null

Freshness

Last updated 2026-05-04 05:30:20; freshness should be verified.

Geography

null