Assessing the consistency of public human tissue RNA-seq data sets

Area of Science:

• Genomics
• Bioinformatics

Background:

• RNA-sequencing (RNA-seq) is widely used for gene expression analysis.
• Concerns exist regarding data comparability across studies due to batch effects and varied methodologies.

Purpose of the Study:

• To assess the consistency of RNA-seq gene expression measurements across different studies.
• To identify factors contributing to systematic differences in RNA-seq data.

Main Methods:

• Compared gene expression data from human brain, heart, and kidney samples across multiple RNA-seq studies.
• Re-analyzed data using a consistent preprocessing pipeline to mitigate bioinformatics bias.

Main Results:

• Published human tissue RNA-seq expression measurements show relative consistency.
• Samples predominantly cluster by tissue type (brain, heart, kidney) rather than by laboratory of origin after basic preprocessing.

Conclusions:

• RNA-seq data from different studies are largely comparable for human tissues.
• Simple preprocessing steps can harmonize data, supporting data fusion and meta-analysis for broader insights.