A novel joint analysis framework improves identification of differentially expressed genes in cross disease transcriptomic analysis

Abstract

MotivationDetecting differentially expressed (DE) genes between disease and normal control group is one of the most common analyses in genome-wide transcriptomic data. Since most studies don’t have a lot of samples, researchers have used meta-analysis to group different datasets for the same disease. Even then, in many cases the statistical power is still not enough. Taking into account the fact that many diseases share the same disease genes, it is desirable to design a statistical framework that can identify diseases’ common and specific DE genes simultaneously to improve the identification power.ResultsWe developed a novel empirical Bayes based mixture model to identify DE genes in specific study by leveraging the shared information across multiple different disease expression data sets. The effectiveness of joint analysis was demonstrated through comprehensive simulation studies and two real data applications. The simulation results showed that our method consistently outperformed single data set analysis and two other meta-analysis methods in identification power. In real data analysis, overall our method demonstrated better identification power in detecting DE genes and prioritized more disease related genes and disease related pathways than single data set analysis. Over 150% more disease related genes are identified by our method in application to Huntington’s disease. We expect that our method would provide researchers a new way of utilizing available data sets from different diseases when sample size of the focused disease is limited.Electronic supplementary materialThe online version of this article (10.1186/s13040-018-0163-y) contains supplementary material, which is available to authorized users.

A Novel Joint Gene Set Analysis Framework Improves Identification of Enriched Pathways in Cross Disease Transcriptomic Analysis

“…The joint analysis methods developed in other fields of omics data analysis have proven useful in increasing the identification power by borrowing information from other similar diseases (Chen X. et al, 2013; Chung et al, 2014; Wang et al, 2016; Lin et al, 2017). In our previous study, we also demonstrated that our joint analysis framework aiming at DE gene detection is more advantageous than single data set analysis and meta-analysis in both simulation studies and real data cases combining different similar disease data sets (Qin and Lu, 2018).…”

“…In this model, Pr( D 1 …, D N ) and α = {α 1 ,α 2 , …α N } need to be estimated from the data. This is a typical mixture model problem, therefore an EM algorithm is implemented to obtain the maximum likelihood estimate of these parameters following the derivation in previous literature (Pounds and Morris, 2003; Qin and Lu, 2018). The details are described as follows:…”

“…Assume that there is a total of 1,000 DE genes out of 10,000 genes. The expression value of each gene in a sample within each data set is generated as described in our previous study (Qin and Lu, 2018) with different means and variance set for each gene. We further assume that the number of data sets to be jointly analyzed is fixed at N = 2 and the number of shared DE genes between two data sets is fixed at 600, 700, 800, or 900, so the DE gene similarity between two data sets are defined as the average shared percentage of DE genes i.e., 12(Pr(D2=1|D1=1)+Pr(D1=1|D2=1)) would be 60, 70, 80, and 90%.…”

“…In this study, we extended our previous joint gene analysis framework to joint gene set analysis framework. Base on the assumption that similar disease tends to share similar disease-related genes and pathways (Carson et al, 2017; Qin and Lu, 2018), we developed two joint gene set analysis frameworks aiming at improving identification power of enriched gene sets by borrowing different levels of information from other similar diseases. Compared with previous joint gene analysis framework, we unified DE gene/pathway statistic modeling through a two-component beta-uniform mixture model of p -values and combined the model with normalized Kolmogorov-Smirnov (KS) statistic for joint gene set enrichment analysis.…”

A Herpes Simplex Virus Thymidine Kinase-Induced Mouse Model of Hepatocellular Carcinoma Associated with Up-Regulated Immune-Inflammatory-Related Signals

“…Computational methods have found increased application in analyzing the complexities underlying experimental studies and have been used in diverse areas of biomedical and cancer research such as to investigate expression profiles [ 12 , 13 ] and genomic alterations [ 14 , 15 ]; the role in miRNA in tumor origin localization [ 16 , 17 ]; in human HCC [ 18 , 19 ]; in the prediction of disease and cancer genes [ 20 , 21 ]; and in cancer prognosis [ 22 , 23 ]. To study the possible causes of HCC, we investigated the transcriptome of HSV-tk mouse liver tissues with HCC and hepatitis using computational analysis of microarray data.…”