Semi-supervised gene shaving method for predicting low variation biological pathways from genome-wide data.

Dongxiao Zhu

doi:10.1186/1471-2105-10-s1-s54

Abstract

BackgroundThe gene shaving algorithm and many other clustering algorithms identify gene clusters showing high variation across samples. However, gene expression in many signaling pathways show only modest and concordant changes that fail to be identified by these methods. The increasingly available signaling pathway prior knowledge provide new opportunity to solve this problem.ResultsWe propose an innovative semi-supervised gene clustering algorithm, where the original gene shaving algorithm was extended and generalized so that prior knowledge of signaling pathways can be incorporated. Different from other methods, our method identifies gene clusters showing concerted and modest expression variation as well as strong expression correlation. Using available pathway gene sets as prior knowledge, whether complete or incomplete, our algorithm is capable of forming tightly regulated gene clusters showing modest variation across samples. We demonstrate the advantages of our algorithm over the original gene shaving algorithm using two microarray data sets. The stability of the gene clusters was accessed using a jackknife approach.ConclusionOur algorithm represents one of the first clustering algorithms that is particularly designed to identify signaling pathways of low and concordant gene expression variation. The discriminating power is achieved by manufacturing a principal component enriched by signaling pathways.

Highlights

The gene shaving algorithm and many other clustering algorithms identify gene clusters showing high variation across samples
We aim to demonstrate that the proposed algorithm is capable of identifying tightly regulated gene sets showing modest and concerted variation using incomplete prior knowledge and real-world microarray data set
Ground truth, which indicates a "complete" gene set used as precondition for applying Gene Set Enrichment Analysis (GSEA) algorithm [14,16], is desirable to demonstrate the claimed advantages of our algorithm

Summary

Introduction

The gene shaving algorithm and many other clustering algorithms identify gene clusters showing high variation across samples. Gene expression in many signaling pathways show only modest and concordant changes that fail to be identified by these methods. Gene clustering that assigns group membership(s) to each gene is a widespread pattern extraction technique. Genes sharing the same membership are often hypothesized to be regulated by the same defined or undefined genomic influence, such as cellular pathway. Model-free clustering techniques such as K-means and hierarchical clustering [1,2,3] are widely used. One limitation of these approaches, as pointed out by many researchers, e.g. [4], is that each gene can only belong to a single cluster.

Objectives

Methods

Results

Discussion

Conclusion