Identifying Subspace Gene Clusters from Microarray Data Using Low-Rank Representation

Yan Cui,Chun-Hou Zheng,Jian Yang,Paolo Provero

doi:10.1371/journal.pone.0059377

Yan Cui, Chun-Hou Zheng + Show 2 more

Open Access

https://doi.org/10.1371/journal.pone.0059377

Copy DOI

Abstract

Identifying subspace gene clusters from the gene expression data is useful for discovering novel functional gene interactions. In this paper, we propose to use low-rank representation (LRR) to identify the subspace gene clusters from microarray data. LRR seeks the lowest-rank representation among all the candidates that can represent the genes as linear combinations of the bases in the dataset. The clusters can be extracted based on the block diagonal representation matrix obtained using LRR, and they can well capture the intrinsic patterns of genes with similar functions. Meanwhile, the parameter of LRR can balance the effect of noise so that the method is capable of extracting useful information from the data with high level of background noise. Compared with traditional methods, our approach can identify genes with similar functions yet without similar expression profiles. Also, it could assign one gene into different clusters. Moreover, our method is robust to the noise and can identify more biologically relevant gene clusters. When applied to three public datasets, the results show that the LRR based method is superior to existing methods for identifying subspace gene clusters.

Highlights

With the advent of the DNA microarray technology, it is possible to study the transcriptional response of a complete genome to different experimental conditions
Low-Rank Representation Before we present the Low-Rank Representation(LRR) based method for identifying gene clusters from microarray data, we first introduce the algorithm of Low-Rank Representation, which is a new framework for seeking the lowest rank representation matrix [27]
The average receiver operator characteristic (ROC) curves are shown in Figure 2 with four different Signal-to-Noise Ratios (SNR)

Summary

Introduction

With the advent of the DNA microarray technology, it is possible to study the transcriptional response of a complete genome to different experimental conditions. Since genes can be clustered with similar cellular functions, identifying the clusters from DNA microarray data might provide much deeper insight into biological function and relevance Traditional clustering methods, such as hierarchical clustering [8], K-means clustering [9], self-organizing maps [10], and modelbased methods [11,12,13,14] can organize gene expression data into clusters of genes possessing similar expression profiles using all the conditions, and the clusters are exclusive and exhaustive (Figure 1 (A)). These algorithms force each gene into a cluster, which may cause the algorithm to be sensitive to noise [15,16,17]

Objectives

Methods

Results

Conclusion