ReGEO

ReGEO restructures Gene Expression Omnibus (GEO) metadata using advanced text mining to extract experimental attributes, including number of time points and disease investigated, to improve the searchability and categorization of GEO series.

Key Features:

• Text mining: Parses unstructured English metadata in GEO series using advanced text mining and sophisticated text analysis.
• Attribute extraction: Automatically extracts the number of time points and the disease investigated from each series' metadata.
• Existing attribute indexing: Retains and indexes existing GEO attributes such as platform organism, experiment type, associated PubMed ID, and keywords from study descriptions.
• Metadata restructuring: Reorganizes and categorizes series-level metadata into structured attributes to enhance query specificity.
• Search enhancement: Improves specificity and searchability of queries across GEO series by leveraging extracted and indexed attributes.

Scientific Applications:

• Dataset discovery: Facilitates more efficient discovery and retrieval of relevant GEO datasets based on structured metadata.
• Big Data analyses: Supports large-scale analyses that require rapid access to structured GEO metadata for many series.
• Time-series and disease-focused selection: Enables selection of time-series experiments and disease-specific datasets by using extracted time-point and disease attributes.

Methodology:

Applies advanced text mining and sophisticated text analysis to parse unstructured GEO series metadata, extract attributes (number of time points and disease), reorganize and categorize series-level metadata, and index platform organism, experiment type, associated PubMed ID, and keywords for search.