Brief Introduction to Some New Results in Gene Expression Analysis, Systems Biology Modeling, Motif Identification, and (noncoding) RNA Analysis

Abstract

This issue of the Journal of Bioinformatics and Computational Biology presents a number of new approaches to several key problems in the analysis of data from biological experiments. These approaches are briefly summarized below. The possibility of using gene expression profiling by microarrays for diagnostic and prognostic purposes has generated much excitement and research for the last 10 years. Nevertheless, a number of issues persist such as how to identify genes that are meaningful in explaining the difference in response and phenotypes, how to rectify batch effects, and how to deal with censoring when modeling survival outcome. In this issue, Zhao and Wang apply correlation principal component regression to handle censoring in survival data under a semi-parametric additive risk model to identify genes which are significantly related to patient survival of a disease. Their technique is shown to be effective on several datasets and is comparatively convenient to implement. The study of genetic regulatory systems has also received a major impetus from microarray technology, which permits the spatial-temporal expression levels of genes to be measured in a massively parallel way. In particular, there is keen interest in inferring regulatory networks from gene expression data. In this issue, Kimura et al. consider a method based on linear programming machines for inferring gene regulation and discuss an approach for assigning confidence values to the inferred gene regulation. This approach is shown to reduce false-positive regulations in a simulated experiment and to produce reasonable confidence values in a real experiment. On the other hand, Fujita et al. attempt to infer Granger causality between sets of genes. They propose a method for identification of Granger causality and a statistical test based on nonparametric bootstrapping. The effectiveness of this approach is demonstrated by simulations and by a successful application to actual biological gene expression data. Metal ions have a major effect on the metabolic processes. For example, zinc (II) ions have been linked to inhibition of Krebs cycle and alteration of energy metabolism. Several distinct mechanisms for zinc (II) inhibition of Krebs cycle have been proposed. In this issue, Čuperlović-Culf constructs a mathematical model of