Type 2 Diabetes Signature Genes Identified via Machine Learning on scRNA-seq Data

Four candidate signature genes for Type 2 diabetes were identified through machine learning analysis of single-cell RNA sequencing data. The dataset includes diagnostic performance metrics and validation results from a clinical cohort of 15 patients and 20 controls. The research was authored by Fang Tang and shared on figshare under a CC-BY-4.0 license.

Use Cases

• Validate potential diagnostic biomarkers for Type 2 diabetes based on the four identified core genes.
• Investigate cell-specific disease mechanisms using the single-cell RNA sequencing data referenced in the study.
• Train machine learning models for disease classification using the gene expression patterns described.
• Explore intercellular signaling networks in diabetes, focusing on the MK and SPP1 pathways highlighted in the analysis.

Strengths

• Includes validation results from an independent clinical cohort of 35 individuals.
• Reports specific diagnostic performance metrics, including AUC values for four candidate genes.
• The underlying analysis integrated differential expression, LASSO regression, and functional enrichment.

Limitations

• The dataset is very small at 25.0 KB, suggesting limited raw data is included.
• Column-level documentation is absent; field semantics must be inferred after download.
• The description does not specify the number of rows or samples in the primary scRNA-seq dataset.

Provenance

Source

figshare

Collection Method

Analysis of scRNA-seq data followed by machine learning (LASSO regression) and clinical validation via qRT-PCR.

Freshness

Last updated 2026-05-12 04:19:07