A Top-Performing Algorithm for the DREAM3 Gene Expression Prediction Challenge

Abstract

A wealth of computational methods has been developed to address problems in systems biology, such as modeling gene expression. However, to objectively evaluate and compare such methods is notoriously difficult. The DREAM (Dialogue on Reverse Engineering Assessments and Methods) project is a community-wide effort to assess the relative strengths and weaknesses of different computational methods for a set of core problems in systems biology. This article presents a top-performing algorithm for one of the challenge problems in the third annual DREAM (DREAM3), namely the gene expression prediction challenge. In this challenge, participants are asked to predict the expression levels of a small set of genes in a yeast deletion strain, given the expression levels of all other genes in the same strain and complete gene expression data for several other yeast strains. I propose a simple -nearest-neighbor (KNN) method to solve this problem. Despite its simplicity, this method works well for this challenge, sharing the “top performer” honor with a much more sophisticated method. I also describe several alternative, simple strategies, including a modified KNN algorithm that further improves the performance of the standard KNN method. The success of these methods suggests that complex methods attempting to integrate multiple data sets do not necessarily lead to better performance than simple yet robust methods. Furthermore, none of these top-performing methods, including the one by a different team, are based on gene regulatory networks, which seems to suggest that accurately modeling gene expression using gene regulatory networks is unfortunately still a difficult task.

Differential expression analysis with global network adjustment

“…In this manner, the predictive weight would be unevenly shared by the correlated transcripts. However, Ruan et al [6] obtained good success in predicting gene expression by using a k-nearest-neighbor (KNN) method. The KNN method used the average of co-expressed genes as a predictor, which gives equal predictive weight to a set of correlated transcripts.…”

“…Some network methods do use extensive prior data such as Ruan et al [6] who used the cluster average of a set of co-expressed genes as a prediction for another gene within the cluster. This relatively simple method had similar or superior accuracy to models that used substantial auxiliary data including regulatory pathways and DNA-binding patterns of transcription factors such as the approach taken by Gustafsson et al [7] that used an “elastic-net" penalty [8] and information from multiple genomic modalities (microarray, ChIP-seq, etc.).…”

“…This relatively simple method had similar or superior accuracy to models that used substantial auxiliary data including regulatory pathways and DNA-binding patterns of transcription factors such as the approach taken by Gustafsson et al [7] that used an “elastic-net" penalty [8] and information from multiple genomic modalities (microarray, ChIP-seq, etc.). Along the lines of Ruan et al [6], we present a simple and computationally efficient method for predicting gene expression. Our method extends the work of Congrove et al [9] who modeled a gene’s expression as linearly dependent on other genes, but we have added features that improve computational efficiency and robustness to variations in transcriptional networks.…”

A hyper-graph approach for analyzing transcriptional networks in breast cancer

“…One of the most significant pieces of information that can potentially be extracted from microarray data is the delineation of transcriptional regulators of biological processes. While many different algorithms have been used to model transcriptional networks from microarray data [20,13], it remains a significant challenge [18].…”

Gene Expression Prediction by Soft Integration and the Elastic Net—Best Performance of the DREAM3 Gene Expression Challenge

“…We also thank Neil Clarke and his collaborators for providing the data for the present challenge and Jianhua Ruan for kindly sending us the manuscript for [11] prior to publication.…”