TopFed: TCGA tailored federated query processing and linking to LOD

Area of Science:

• Bioinformatics
• Genomics
• Data Science

Background:

• The Cancer Genome Atlas (TCGA) project aims to catalog cancer-causing genetic mutations.
• Analyzing large TCGA datasets is challenging due to data size and parsing requirements.
• Virtual data integration for TCGA analysis is hindered by data retrieval difficulties.

Purpose of the Study:

• To transform TCGA data into RDF format for Semantic Web integration.
• To enable efficient querying of distributed TCGA data.
• To develop a federated SPARQL query engine tailored for TCGA.

Main Methods:

• TCGA data converted to Resource Description Format (RDF).
• Linked to Linked Open Data (LOD) cloud datasets.
• Data distributed across SPARQL endpoints with a federated query engine (TopFed).

Main Results:

• TopFed demonstrated superior performance compared to FedX.
• TopFed selected fewer data sources with 100% recall.
• Query execution time was reduced to one-third of FedX's.

Conclusions:

• TopFed provides a single-point-of-access for distributed TCGA data.
• The system facilitates effective biomedical research with large-scale TCGA data.
• TopFed overcomes limitations of local resources for data retrieval and parsing.