gwSPIA: Improved Signaling Pathway Impact Analysis With Gene Weights

Abstract

Gene set analysis using signaling pathway has become a popular downstream analysis following differential expression analysis. From a biological point of view, only some portions of a pathway are expected to be altered; however, a few approaches using the different importance of genes in signaling pathways, which encompass the constitutive functional nonequivalent roles of genes in real pathways, have been proposed and none of them tries to associate the importance of genes with the related disease. In this paper, we developed an extended method of signaling pathway impact analysis (SPIA), called gwSPIA, by incorporating three signaling pathway-based gene weight merits that reflect the importance of genes from different aspects and attempt to associate the importance of genes with the related diseases. By applying the gwSPIA to the gene expression data sets in comparison with other seven methods in three measures, sensitivity, prioritization, and specificity, we show that the gwSPIA ranks in the second place in both sensitivity and prioritization. Furthermore, the specificity of the gwSPIA is better than SPIA, which is lower than 25%. The results also suggest that the gene weight used in the gwSPIA can reflect the association between the genes and the related diseases. The R package of the gwSPIA can be accessed from https://github.com/sterding/gwSPIA. INDEX TERMS Differentially expressed genes, gene weights, gwSPIA, signaling pathways analysis.

An Entropy-Based Directed Random Walk for Cancer Classification Using Gene Expression Data Based on Bi-Random Walk on Two Separated Networks

“…In cancer classification, a robust gene weight merit is vital to reflect the importance of genes from different aspects and establish significant genes with related diseases [ 10 ]. Several studies in pathway-based methods typically use the t -test as the gene weighting method to measure the gene expression levels for further cancer classification.…”

“…Consider that those existing pathway-based methods including directed random walk (DRW), significant DRW, and pathway activity inference using condition-responsive genes (PAC method) all targeted on the t -test as the single statistical measurement to weigh each gene in the gene expression data. However, the lack of a comprehensive gene weighting method could affect the classification performance of pathway-based methods [ 10 , 11 ].…”

Regularized Sparse Modelling for Microarray Missing Value Estimation

“…It has been known to us that DNA microarray technology provides researchers a high-throughput way to efficiently obtain the gene expression levels of a certain disease from different environments, subjects, tissues, and cell cycles and that microarray data analysis greatly facilitates the identification of disease genes and the diagnosis of cancers and tumor subtypes [1,2]. Accordingly, researchers have utilized a wealth of statistical analysis and machine learning models (e.g., classification, clustering, feature selection, network analysis, and causal inference) to analyze gene expression profiles towards understanding the underlying biological mechanisms [3,4]. However, both human and non-human factors, including, but not limited to, false positive PCR, inappropriate use of test chips, impurity of chip surface, and insufficient resolution of fluorescent images, can result in gene expression profiles with missing entries [5].…”

Signaling Pathway Analysis Combined With the Strength Variations of Interactions Between Genes Under Different Conditions

“…Functional Link Enrichment of Gene Ontology or gene sets (LEGO) considered a network-based gene weight of genes and an ORA method [22]. The method gwSPIA was also a gene-weight-based method which improved SPIA with three different gene weights: impact factor (IF), betweenness centrality (BC), and specificity (SP) [23]. It takes the complex interactions between genes and the importance of genes which can reflect the association between genes and diseases in consideration at the same time [23].…”