Gene Expression Signature for Distinguishing MASL and MASH Liver Disease

A 2026 study by Guo Wu Lin integrates five datasets from the Gene Expression Omnibus to identify a gene signature for liver disease. The training cohort comprises 149 MASL and 158 MASH samples, validated on a separate cohort of 51 MASL and 155 MASH samples. It uses machine learning to derive a six-gene diagnostic model validated with clinical samples.

Use Cases

• Training a diagnostic classifier for metabolic liver disease based on the identified six-gene signature.
• Analyzing immune cell infiltration patterns in liver disease based on the CIBERSORT analysis mentioned.
• Investigating cell-type-specific gene expression in liver disease using the single-cell RNA sequencing data referenced.
• Benchmarking machine learning feature selection methods for transcriptomic data based on the LASSO, SVM-RFE, and random forest approach described.

Strengths

• Model validation on an independent cohort of 206 samples (51 MASL, 155 MASH).
• Diagnostic model achieved an AUC of 0.893 in the validation cohort.
• Includes validation with quantitative PCR on 60 clinical liver tissue samples.
• Integrated analysis from five public GEO datasets.

Limitations

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

Provenance

Source

Gene Expression Omnibus (GEO) datasets, integrated by the study author.

Collection Method

Integrated transcriptomic analysis, machine learning feature selection, and clinical PCR validation.

Time Range

null

Freshness

Last updated 2026-04-20 04:11:46; freshness should be verified.

Geography

null