Computational insights on the molecular mechanisms across breast cancer progression combining gene differential expression and co-expression

Abstract

As a complex disease with mechanisms that are not fully understood, breast cancer pathology and progression is expected to be determined not only by individual genes, but also by their coordinated effects though a more systemic framework. In this contribution, we combine gene differential analysis (single gene view) with gene co-expression analysis (systemic view) to provide insights on the implicated molecular mechanisms across breast cancer progression. Important gene-gene links in a gene co-expression network are identified and clustered with Density and Distance Content driven (DeDiCo) algorithm, which has been recently published by members of our groups. This algorithm presents very good performance regarding its accuracy on both synthetic and real-life datasets and ability to determine arbitrarily shaped clusters without any assumptions on the shape and the number of the clusters. This work provides a pipeline, exploiting a novel clustering algorithm, in order to analyse gene-gene links based on their co-expression relationships and gene differential analysis across breast cancer progression stages. Existing bibliography verifies the validity of the resulted clusters while exclusive pathways per breast cancer stage as well as a common pathway signature across all breast cancer stages are reported and proper discussion takes place.