Public Omics Explorer

Public Omics Explorer performs literature-informed dataset retrieval, semantically linking GEO datasets and ENA records through their associated PubMed publications to enable integrative semantic search across GEO omics datasets.

Key Features:

• Automated Data Collection: POE automatically collects and indexes GEO metadata, ENA cross-references, and PubMed abstracts on a daily basis.
• Semantic Embedding: POE employs the biomedical-specialized SBioBERT model to generate dense vector representations from publication text.
• Similarity Search Indexing: Embeddings are indexed using Facebook AI Similarity Search (FAISS) to enable high-precision, context-aware retrieval.
• Natural Language Search: Users can search using free-text natural language queries, which are processed through semantic search to identify conceptually relevant datasets based on linked publication content.
• Structured Filtering: Search results can be refined by organism, experiment type, library strategy, sample type, extracted molecule, and publication year.
• Accession-Based Retrieval: POE supports direct retrieval of datasets via accession identifiers including GSE IDs, PubMed IDs, and DOIs.
• Programmatic API: POE offers a RESTful API for integration into computational pipelines and automated workflows.
• Cross-Database Linking: POE links processed data in GEO with raw data in ENA through shared publication context.

Scientific Applications:

• Hypothesis Generation: POE facilitates hypothesis generation by enabling discovery of relevant datasets through publication-based semantic relationships.
• Meta-Analysis: POE supports meta-analysis by identifying and retrieving related omics datasets across GEO and ENA through their shared literature context.
• Exploratory Research: POE enables exploratory research by providing context-aware dataset discovery beyond keyword matching.

Methodology:

POE employs SBioBERT for semantic embedding of PubMed publication text, with dense vector representations indexed using FAISS for similarity search. Natural language queries are processed through this semantic search infrastructure to retrieve conceptually relevant GEO datasets and ENA records linked through shared publications.