BioDataome: a collection of uniformly preprocessed and automatically annotated datasets for data-driven biology

Area of Science:

• Bioinformatics
• Genomics
• Biotechnology

Background:

• The rapid expansion of omics data necessitates efficient organization and curation.
• Publicly available omics datasets often lack uniform preprocessing and disease annotation, hindering reuse.
• Biomedical knowledge organization into accessible online databases is crucial for research advancement.

Purpose of the Study:

• To develop BioDataome, a database for uniformly preprocessed and disease-annotated omics data.
• To facilitate the reuse of public omics data for large-scale experiments and meta-analyses.
• To provide a user-friendly web application and R package for data querying and downloading.

Main Methods:

• Implemented a standardized preprocessing pipeline for microarray gene expression, RNA-Seq gene expression, and DNA methylation data.
• Automated disease ontology term annotation for all datasets.
• Developed methods to identify datasets with common samples and automatically discover control samples in case-control studies.

Main Results:

• BioDataome currently houses approximately 5600 datasets and 260,000 samples across nearly 500 diseases.
• All datasets are uniformly processed, disease-annotated, and readily available for downstream analysis.
• The database supports large-scale experiments and meta-analyses, with demonstrated utility through exploratory data analysis examples.

Conclusions:

• BioDataome significantly enhances the accessibility and usability of public omics data.
• The database promotes accelerated biomedical research by providing high-quality, curated datasets.
• An accompanying R package further supports researchers in leveraging BioDataome resources.