Fig 1.
The process consists of five main stages. The main tasks in each stage are described in the corresponding box.
Table 1.
The nine public gene expression datasets included in our meta-analysis.
Fig 2.
Principal component analysis of the combined dataset after batch effect correction.
All samples are represented by different symbols with shapes according to their tissue origins and colors based on their experimental conditions.
Fig 3.
The fold changes of top 14 immune genes across different datasets.
The small negative fold changes of Mrc1 and Hmox1 genes in the dataset GSE41941 (containing two replicates) are likely due to study-specific bias.
Fig 4.
The heatmap of DE genes identified during meta-analysis.
A total of 691 DE genes were identified (394 up-regulated and 297 down-regulated). The horizontal lines on the right indicate those genes belong to several immune related functional groups.
Table 2.
The genes are ranked by their effect sizes and separated based on their directions of changes.
Fig 5.
PPI subnetwork analysis of DE genes.
(A) The subnetwork using all DE genes. Nodes in red indicate upregulated genes, nodes in green indicate down-regulated genes, and nodes in grey are interaction partners of the DE genes; (B) The PPI subnetwork for up-regulated genes; (C) The PPI subnetwork for down-regulated genes.