A systems biology approach to identify proliferative biomarkers and pathways in breast cancer

Proliferation is the most fundamental hallmark of cancer. In breast cancer, proliferation is assessed by staining cells for the nuclear antigen Ki-67. Ki-67 has a potential use as a prognostic and predictive factor for therapy in breast cancer. In this study, an Artificial neural network (ANN) using a feedforward multi-layer perceptron back propagation algorithm was applied to facilitate a systems biology viewpoint to identify genes associated with Ki-67 in 3 different breast cancer gene expression studies. Subsequently, a non-reductionist ANN based network was created for the top ten STRING identified genes for Ki-67 to identify the cross linked hubs (UBE2C, KIF2C) in the map. The genes which were found in common among the top 200 probes in the three cohorts and as cross links in the map were hypothesized to be more powerful and found to be significantly associated with prognosis in breast cancer. In future, this work can be expanded to include other proliferation associated targets to map the system in a more detailed way and see the influence of the common markers obtained in drug resistance (endocrine resistance) in breast cancer. The network map constructed, represents a novel non-recursive additive approach to visualize a gene network and provides a dynamic view for the disease of interest.