Identifying cooperative genes causing cancer progression with dynamic causal inference

Abstract

It is well known that some gene aberrations can cause cancer by disrupting the delicate balance of critical biological processes at the cellular level. Such aberrations are rare and are not limited to gene mutations alone and hence are difficult to be identified from data. Moreover, focusing exclusively on gene aberrations neglects other significant aspects of cancer development such as the fact that cancer occurs due to gene interactions evolving as a dynamical system. Therefore, expanding our knowledge about the dynamics of genetic mechanisms that cause cancer is crucial for a comprehensive understanding of cancer development. In this paper, a novel causal method for identifying collaborative networks of cancer drivers based on dynamic system analysis is introduced. The method integrates the temporal dimension of the data throughout cancer progression and provides a way of testing for the causality of candidate genes in cancer. We have applied our method to single-cell and bulk sequencing datasets of breast cancer. The evaluation results show that our method systematically identifiesbona fidedriver genes and detects sets of genes strongly linked to cancer progression. The results suggest that our method can discover mutated and non mutated drivers of cancer to provide a comprehensive view of cancer development.R package implementing our approach as well as scripts for the experiments and datasets used can be found athttps://github.com/AndresMCB/DynamicCancerDriverKM.