Project Summary/Abstract The overall objective of this proposal is to develop general analytic frameworks to characterize patient-specific pathway signatures by sequential estimations of cancer-specific (global) and patient-specific (local) networks. We develop innovative and flexible knowledge-guided quantitative frameworks that integrate multiple sources of information: qualitative and unstructured knowledge databases, data-driven de novo causal structures as well as upstream multi- platform molecular profiling data at the genomic, epigenomic, transcriptomic and proteomic levels. Our methods are motivated by and applied to novel, unpublished, reverse-phase protein array-based proteomic profiles generated on patient tumor samples across 32 cancer types from The Cancer Genome Atlas (TCGA) as well as cell line samples from the MD Anderson Cell lines Project (MCLP) across 19 tumor lineages. This allows us to comprehensively characterize commonalities and differences in network biology across tumor lineages to provide insight into the underlying biological mechanisms, and discovery of reliable unsupervised and supervised prediction models for relevant clinical and drug sensitivity outcomes -- to aid translational and precision medicine.

Narrative This project will have significant impact on integrated network analysis to characterize patient-specific pathway signatures by sequential estimations of cancer-specific and patient-specific networks. We utilize vast amount of data at genomic, epigenomic, transcriptomic, and proteomic levels on patient tumor samples from 32 tumor types and tumor-driven cell line samples (with drug sensitivity data) from 19 tumor lineages. This study will bring comprehensive picture of pathway activity across patients from different tumor lineages and help achieve our goal of devising effective treatment and prevention strategies for cancer via statistical integrated network approaches.