LURE

LURE identifies novel cancer driver events in TCGA and other multidimensional genomic datasets by associating gene expression signatures with mutation events using a semi-supervised learning framework.

Key Features:

• Semi-Supervised Learning: Employs a progressive label learning framework to iteratively refine predictions by leveraging both labeled (known) and unlabeled samples.
• Minimum Spanning Analysis: Uses minimum spanning analysis to identify connections between gene expression signatures across different samples.
• Expression–Mutation Association: Predicts cancer drivers by associating altered samples with known mutation events based on shared gene expression signatures.
• Large-Scale Dataset Application: Designed to operate on large-scale cancer genome projects such as The Cancer Genome Atlas (TCGA).

Scientific Applications:

• TCGA Analysis: Applied to TCGA datasets to identify high-confidence relationships between 53 novel and 18 known mutation events.
• Pathway Insights: Revealed connections involving TP53, telomere maintenance, and MAPK/RTK signaling pathways.
• Genetic Interaction Discovery: Uncovered alterations within the same gene, family, or pathway, indicating complex genetic interactions relevant to candidate targeted therapies.

Methodology:

Implements a progressive semi-supervised (label) learning framework together with minimum spanning analysis to associate gene expression signatures with known mutation events.