Quantifying Gene Co-Expression Heterogeneity in Cancer Towards Efficient Network Biomarker Design
It is well known that cancer is a highly heterogeneous disease, and the predictive capability
of targeted gene signature approach suffers from the inter-tumor heterogeneity. Here we propose a
framework to quantify the molecular heterogeneity of tumors from gene-gene relational perspective
using co-expression networks and interactome data. We believe that to understand individualized gene
behavior across patients, relational status of genes needs to be considered because complex disease phenotype is often
caused by failures of genetic interactions in cancer cells.
We quantified gene-gene relational heterogeneity from a benchmark data set using co-expression networks inferred from
Microarray data, and showed that genes related to breast cancer metastasis can be stratified to different classes based on
their relational status obtained from pair-wise comparisons of co-expression networks. Further we used the relational
heterogeneity information to predict patient survival and found that relationally heterogeneous gene set is less predictive
than relatively conserved cancer genes. We explored heterogeneity gene sets using interactome data and identified densely
connected components that are causal to inter-tumor heterogeneity. We independently validated our approach with two
patient cohorts. Our results demonstrated the efficiency of using heterogeneity information to design network markers.
Keywords: Biomarkers, breast cancer, co-expression network, network biology, tumor heterogeneity.
Rights & PermissionsPrintExport