Hierarchy of Gene Expression Data is Predictive of Future Breast Cancer Outcome

We calculate measures of hierarchy in gene and tissue networks of breast cancer patients. We find that the likelihood of metastasis in the future is correlated with in-creased values of network hierarchy for expression networks of cancer-associated genes, due to correlated expression of cancer-specific pathways. Conversely, future metasta-sis and quick relapse times are negatively correlated with values of network hierarchy in the expression network of all genes, due to dedifferentiation of gene pathways and circuits. These results suggest that hierarchy of gene expression may be useful as an additional biomarker for breast cancer prognosis.