Massive mining of publicly available RNA-seq data from human and mouse

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• RNA sequencing (RNA-seq) is crucial for genome-wide transcript quantification.
• Raw RNA-seq data presents a significant barrier for large-scale retrospective analyses.
• A centralized, processed resource is needed to facilitate global transcriptomic studies.

Purpose of the Study:

• To develop ARCHS4, a web resource providing processed RNA-seq data for human and mouse.
• To make the majority of published RNA-seq data accessible at gene and transcript levels.
• To enable intuitive exploration and analysis of large-scale transcriptomic datasets.

Main Methods:

• Utilized cloud infrastructure to align FASTQ files from the Gene Expression Omnibus (GEO).
• Processed and curated a large collection of human and mouse RNA-seq samples.
• Developed a web interface for data querying, visualization, and gene-centric information retrieval.

Main Results:

• ARCHS4 hosts 187,946 samples (103,083 mouse, 84,863 human).
• Processed data is available at gene and transcript levels.
• Web interface offers average expression, co-expressed genes, and predicted functions/interactions.

Conclusions:

• ARCHS4 significantly lowers the barrier for global and integrative retrospective RNA-seq analyses.
• The resource facilitates deeper understanding of gene expression patterns and biological functions.
• ARCHS4 serves as a valuable tool for researchers in genomics and transcriptomics.