Incorporating higher-order representative features improves prediction in network-based cancer prognosis analysis

Shuangge Ma; Michael R Kosorok; Jian Huang; Ying Dai

doi:10.1186/1755-8794-4-5

Back

Incorporating higher-order representative features improves prediction in network-based cancer prognosis analysis

Journal article

Open access

Peer reviewed

Incorporating higher-order representative features improves prediction in network-based cancer prognosis analysis

Shuangge Ma, Michael R Kosorok, Jian Huang and Ying Dai

BMC medical genomics, Vol.4(1), pp.5-5

01/12/2011

DOI: 10.1186/1755-8794-4-5

PMCID: PMC3037289

PMID: 21226928

Files and links (1)

url

https://doi.org/10.1186/1755-8794-4-5View

Published (Version of record) Open Access

Abstract

In cancer prognosis studies with gene expression measurements, an important goal is to construct gene signatures with predictive power. In this study, we describe the coordination among genes using the weighted coexpression network, where nodes represent genes and nodes are connected if the corresponding genes have similar expression patterns across samples. There are subsets of nodes, called modules, that are tightly connected to each other. In several published studies, it has been suggested that the first principal components of individual modules, also referred to as "eigengenes", may sufficiently represent the corresponding modules. In this article, we refer to principal components and their functions as representative features". We investigate higher-order representative features, which include the principal components other than the first ones and second order terms (quadratics and interactions). Two gradient thresholding methods are adopted for regularized estimation and feature selection. Analysis of six prognosis studies on lymphoma and breast cancer shows that incorporating higher-order representative features improves prediction performance over using eigengenes only. Simulation study further shows that prediction performance can be less satisfactory if the representative feature set is not properly chosen. This study introduces multiple ways of defining the representative features and effective thresholding regularized estimation approaches. It provides convincing evidence that the higher-order representative features may have important implications for the prediction of cancer prognosis.

Predictive Value of Tests

Prognosis

Oligonucleotide Array Sequence Analysis

Humans

DNA, Complementary - genetics

Gene Expression Profiling

Neoplasms - diagnosis

Models, Statistical

Algorithms

Neoplasms - genetics

Female

Neoplasms - pathology

Principal Component Analysis

Neural Networks (Computer)

Details

Title: Subtitle: Incorporating higher-order representative features improves prediction in network-based cancer prognosis analysis
Creators: Shuangge Ma - School of Public Health, Yale University, New Haven, CT, USA. shuangge.ma@yale.edu
Michael R Kosorok
Jian Huang
Ying Dai
Resource Type: Journal article
Publication Details: BMC medical genomics, Vol.4(1), pp.5-5
DOI: 10.1186/1755-8794-4-5
PMID: 21226928
PMCID: PMC3037289
NLM abbreviation: BMC Med Genomics
ISSN: 1755-8794
eISSN: 1755-8794
Publisher: England
Grant note: P01 CA142538 / NCI NIH HHS CA120988 / NCI NIH HHS P01 CA142538-02 / NCI NIH HHS P01 CA142538-01 / NCI NIH HHS P30 ES010126 / NIEHS NIH HHS CA142774 / NCI NIH HHS LM009754 / NLM NIH HHS CA142538 / NCI NIH HHS
Language: English
Date published: 01/12/2011
Academic Unit: Statistics and Actuarial Science
Record Identifier: 9983985816202771

Metrics

26 Record Views

19 Times Cited - Web of Science