A model for gene deregulation detection using expression data

TL;DR

This paper introduces a statistical model using an EM algorithm and message passing to detect gene deregulation in cancer cells from expression data, even when differential expression is not apparent.

Contribution

It presents a novel EM-based method that models gene deregulation with hidden states, improving detection of regulatory alterations in cancer subtypes.

Findings

Effective in identifying deregulated genes in simulated data

Successfully applied to bladder cancer dataset

Provides posterior probabilities of gene deregulation

Abstract

In tumoral cells, gene regulation mechanisms are severely altered, and these modifications in the regulations may be characteristic of different subtypes of cancer. However, these alterations do not necessarily induce differential expressions between the subtypes. To answer this question, we propose a statistical methodology to identify the misregulated genes given a reference network and gene expression data. Our model is based on a regulatory process in which all genes are allowed to be deregulated. We derive an EM algorithm where the hidden variables correspond to the status (under/over/normally expressed) of the genes and where the E-step is solved thanks to a message passing algorithm. Our procedure provides posterior probabilities of deregulation in a given sample for each gene. We assess the performance of our method by numerical experiments on simulations and on a bladder cancer data set.