Evaluation of network-guided random forest for disease gene discovery

TL;DR

This study evaluates a network-guided random forest approach for disease gene discovery, finding it improves gene identification when disease genes form modules but does not enhance overall disease prediction accuracy.

Contribution

Introduces and empirically assesses a network-guided RF method that incorporates gene network information into feature sampling for disease gene discovery.

Findings

Network-guided RF better identifies disease genes in modules.

It does not improve overall disease prediction accuracy.

Spurious gene selection occurs with network info when disease status is independent of genes.

Abstract

Gene network information is believed to be beneficial for disease module and pathway identification, but has not been explicitly utilized in the standard random forest (RF) algorithm for gene expression data analysis. We investigate the performance of a network-guided RF where the network information is summarized into a sampling probability of predictor variables which is further used in the construction of the RF. Our results suggest that network-guided RF does not provide better disease prediction than the standard RF. In terms of disease gene discovery, if disease genes form module(s), network-guided RF identifies them more accurately. In addition, when disease status is independent from genes in the given network, spurious gene selection results can occur when using network information, especially on hub genes. Our empirical analysis on two balanced microarray and RNA-Seq breast cancer datasets from The Cancer Genome Atlas (TCGA) for classification of progesterone receptor (PR) status also demonstrates that network-guided RF can identify genes from PGR-related pathways, which leads to a better connected module of identified genes.