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

Conclusion and Discussions for details). MAPE_Gene performs even worse in this scenario because there is not shared DE genes within a pathway. Meta-analysis by taking maximum p-value would thus produce many false positives in DE gene detection. Other methods based on pathway-level evidence summarization remain same performance as in Scenario 1. To sum up, the simulation test with different parameter setup and two different scenarios demonstrates that JointNormKS performs best among all other methods even when there are no shared DE genes within an enriched pathway. We then decide to use JointNormKS method in real data application in next section. Comparison of JointNormKS With Single Data Set Analysis in Real Data Application Based on the simulation test results, we apply the JointNormKS framework on two real data sets and compare their identified enriched gene sets with those derived from single data set analysis, respectively. We use lung and colorectal adenocarcinoma as one example because adenocarcinoma both develop from gland cells of different tissues and as shown in our previous study, we observed that lung and colorectal adenocarcinoma shared a significant higher percentage of DE genes than other cancers (Qin and Lu, 2018). Alzheimer's disease and Huntington's disease are selected as another example due to their highly similar clinical phenotypes. Real Data Application: Lung Adenocarcinoma and Colorectal Adenocarcinoma JointNormKS is first applied...
Source: Frontiers in Genetics - Category: Genetics & Stem Cells Source Type: research