An empirical Bayes approach to inferring large-scale gene association networks

An empirical Bayes approach to inferring large-scale gene association networks

Vol. 21 no. 6 2005, pages 754–764 | Juliane Schäfer and Korbinian Strimmer
This paper presents an empirical Bayes approach to inferring large-scale gene association networks using graphical Gaussian models (GGMs) from gene expression data. The authors address the challenge of inferring network structures in microarray analysis, where the number of genes often exceeds the number of samples, making standard algorithms for graphical models inapplicable. They propose a novel framework that includes improved small-sample point estimates of partial correlation, an exact test of edge inclusion with adaptive estimation of the degree of freedom, and a heuristic network search based on false discovery rate multiple testing. The approach is validated through computer simulations and applied to a breast cancer dataset, demonstrating its ability to recover the true network topology with high accuracy even for small-sample datasets. The method is implemented in the R package 'GeneTS' and is freely available. The authors discuss the advantages and potential drawbacks of their framework, highlighting its practical utility in large-scale gene association network inference.This paper presents an empirical Bayes approach to inferring large-scale gene association networks using graphical Gaussian models (GGMs) from gene expression data. The authors address the challenge of inferring network structures in microarray analysis, where the number of genes often exceeds the number of samples, making standard algorithms for graphical models inapplicable. They propose a novel framework that includes improved small-sample point estimates of partial correlation, an exact test of edge inclusion with adaptive estimation of the degree of freedom, and a heuristic network search based on false discovery rate multiple testing. The approach is validated through computer simulations and applied to a breast cancer dataset, demonstrating its ability to recover the true network topology with high accuracy even for small-sample datasets. The method is implemented in the R package 'GeneTS' and is freely available. The authors discuss the advantages and potential drawbacks of their framework, highlighting its practical utility in large-scale gene association network inference.
Reach us at info@study.space
[slides and audio] An empirical Bayes approach to inferring large-scale gene association networks