Topological data analysis identifies PURPL as a prognostic gene in breast cancer Luminal A patients

Luminal A is the most common breast cancer molecular subtype and it is associated with favorable five-year prognosis. Many patients, however, relapse within 10 years from the time of the initial diagnosis, suggesting underlying heterogeneity of the disease. Using topological data analysis, we previously identified a copy number change in the region 5p14.3-p12 associated to Luminal A patients. In this paper, we further investigate this region and identify the gene p53 Upregulated Regulator of p53 Levels ( PURPL ), a lncRNA not previously associated with breast cancer. We show that PURPL amplification and overexpression are associated with poor survival in Luminal A patients. To study possible mechanisms of action of PURPL , we test whether the low survival of patients with overexpression of PURPL is observed when the data are stratified according to the mutation status of relevant genes or pathways. We find that while the survival differences remain in the subset of patients with no mutations in either TP53 or PIK3CA , they disappear when mutations in both TP53 and PIK3CA are present. These results suggest that the PURPL transcript interacts with the products of both TP53 and PIK3CA and that these interactions determine the different survival outcomes of Luminal A patients with high expression of PURPL .