Unifying cancer and normal RNA sequencing data from different sources.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Next-generation sequencing (NGS) has generated vast amounts of whole transcriptome sequencing (RNA-seq) data from large-scale studies like GTEx and TCGA.
• Comparing RNA-seq data across different studies is challenging due to variations in sample and data processing.
• Decoding complex human diseases requires integrated analysis of diverse RNA-seq datasets.

Purpose of the Study:

• To develop a computational pipeline for processing and unifying RNA-seq data from multiple sources.
• To address the challenges of data heterogeneity and batch effects in large-scale RNA-seq studies.
• To enable reliable comparative analyses between datasets from different origins, such as GTEx and TCGA.

Main Methods:

• Developed a pipeline incorporating uniform realignment and gene expression quantification.
• Implemented batch effect removal strategies beyond standard alignment and quantification.
• Processed and normalized RNA-seq data from the Genotype Tissue Expression (GTEx) project and The Cancer Genome Atlas (TCGA).

Main Results:

• Uniform alignment and quantification alone are insufficient for combining RNA-seq data from different sources.
• Removal of additional batch effects is crucial for facilitating meaningful data comparison.
• Successfully corrected study-specific biases in GTEx and TCGA data, enabling cross-study comparisons.

Conclusions:

• The developed pipeline effectively unifies and normalizes heterogeneous RNA-seq data.
• Batch effect correction is essential for integrating large-scale RNA-seq datasets for disease research.
• Normalized datasets are publicly available to facilitate further research and comparative analysis.