pLasso

pLasso integrates prior biological pathway information into a Bayesian Lasso framework to improve reconstruction of gene networks from genomic data.

Key Features:

• Integration of Prior Knowledge: pLasso incorporates pathway information from Pathway Commons (PC) and the Kyoto Encyclopedia of Genes and Genomes (KEGG) into the network reconstruction process.
• Bayesian Lasso Methodology: It implements a Bayesian version of the Lasso that partitions gene edges into two subsets—edges present in known pathways and edges without prior information—and assigns different prior distributions to these subsets using a modified Bayesian Information Criterion.
• Enhanced Network Recovery: Simulation studies demonstrate that pLasso recovers underlying gene networks more effectively than traditional Lasso methods when leveraging prior biological knowledge.
• Application to Genomic Data: Applied to microarray gene expression datasets, pLasso identified network hub genes associated with clinical outcomes in cancer patients.

Scientific Applications:

• Gene Network Reconstruction: Reconstructing gene networks from large-scale genomic data, including microarray gene expression profiles.
• Pathway Analysis: Incorporating pathway information to elucidate biological processes underlying complex phenotypes.
• Clinical Research: Identifying network hub genes associated with clinical outcomes, as demonstrated in cancer studies.

Methodology:

Partitioning gene edges based on prior pathway information, assigning distinct prior distributions to these partitions within a Bayesian framework, and applying a modified Bayesian Information Criterion to optimize network reconstruction.