Modular RNA-seq Analytics for Exploratory Biomarker Discovery using Public Data

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Publicly available RNA sequencing (RNA-seq) data offer a valuable resource for biomarker discovery.
• Data heterogeneity across studies poses significant challenges for accurate analysis.
• Integrating disparate datasets is crucial for robust biomarker identification.

Purpose of the Study:

• To present a modular analytics pipeline for unifying heterogeneous RNA-seq datasets.
• To standardize quality control, differential expression, and pathway analysis.
• To leverage machine learning for robust biomarker discovery from public data.

Main Methods:

• Developed a modular analytics pipeline integrating open-source tools.
• Standardized quality control, differential expression, and pathway analysis workflows.
• Applied competitive machine learning to merge disparate RNA-seq datasets.

Main Results:

• Demonstrated the pipeline's utility across three disease contexts: COVID-19 severity, sepsis, and atherosclerosis.
• Identified differentially expressed gene signatures for COVID-19 severity.
• Generated concise biomarker panels for sepsis using integrated analysis.
• Examined tissue and cell-type specificity of N-acyl-phosphatidylethanolamine phospholipase D in atherosclerosis.

Conclusions:

• The adaptable, modular pipeline effectively repurposes public RNA-seq data.
• Open-source tools and machine learning reduce noise and generate novel biological hypotheses.
• This approach establishes a foundation for exploratory biomarker discovery using public resources.