Merging microarray studies to identify a common gene expression signature to several structural heart diseases

Area of Science:

• Genomics
• Cardiovascular Medicine
• Bioinformatics

Background:

• Heart disease is a leading global cause of mortality.
• Gene expression signatures offer potential for enhanced diagnosis and treatment of heart disease.
• Publicly available microarray data presents opportunities but also challenges in merging datasets and identifying differentially expressed genes.

Purpose of the Study:

• To identify a robust gene expression signature for various heart diseases.
• To overcome challenges in merging multiple microarray datasets.
• To establish a reliable method for identifying differentially expressed genes in cardiac conditions.

Main Methods:

• Merged nine public microarray datasets, creating a combined dataset of 689 samples and 8354 features.
• Determined adjusted p-values and fold changes, generating 12 sets of differentially expressed genes using various thresholds.
• Employed the random forest algorithm to evaluate classification accuracy for each gene set.

Main Results:

• A set of 62 differentially expressed genes was identified as a potential heart disease signature.
• This signature demonstrated approximately 95% accuracy in classifying heart disease samples.
• The identified genes are implicated in the pathophysiology of heart conditions.

Conclusions:

• A common gene expression signature for different cardiac diseases was successfully identified.
• The methodology is effective for identifying gene expression signatures and applicable to other diseases.
• This finding supports the development of more accurate diagnostic and therapeutic strategies for heart disease.