Gene Expression Signature for Early Discrimination of MASL and MASH Liver Disease

A 2026 study by Guo Wu Lin integrates transcriptomic data from multiple Gene Expression Omnibus cohorts to identify a diagnostic gene signature. The dataset contains results from an analysis of 149 MASL and 158 MASH training samples, plus 51 MASL and 155 MASH validation samples. It includes six key genes identified via machine learning for distinguishing liver disease subtypes.

Use Cases

• Training a classifier for liver disease subtype discrimination based on the six-gene signature (MMP9, FABP5, TREM2, CTSD, UBD, MAP2K1).
• Analyzing immune cell infiltration patterns in liver disease, as the description references CIBERSORT estimates for monocytes and macrophages.
• Validating single-cell RNA sequencing findings against bulk transcriptomic data, referencing analysis of dataset GSE136103.
• Benchmarking artificial neural network performance against other ML algorithms (LASSO, SVM-RFE, Random Forest) for biomarker selection.

Strengths

• Validation in an independent cohort achieved an AUC of 0.893, indicating high diagnostic accuracy.
• Analysis integrates data from five training datasets and one validation dataset, providing a multi-cohort perspective.
• Key findings were validated via quantitative PCR in 60 clinical liver tissue samples.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment for large-scale model training.
• The dataset is very small (73.7 KB), suggesting it contains summary results rather than raw expression data.

Provenance

Source

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

Collection Method

Integrated transcriptomic analysis, differential expression, weighted gene co-expression networks, and machine learning feature selection.

Time Range

Study published in 2026; underlying GEO dataset collection dates are not specified.

Freshness

Last updated 2026-04-20 04:11:44; freshness should be verified.

Geography

null