14 Pathway (Gene Set) Analyses

rnaseqDRaMA performs QuSAGE pathway analysis for all genes in each experimental contrast. The results are added to the input rds file and is visualized by Shiny. In the current release of rnaseqDRaMA (v1.4.0) analyzed gene sets are taken from the c2 section (curated gene sets) of the Molecular Signature Database maintained by the Broad Institute. Specifically, the canonical pathways section contains gene sets from several pathway databases including: KEGG, BIOCARTA, REACTOME, PID, biomedical literature, and domain experts. Specific database can be selected using Plot Pathways from Specific Database dropdown menu in Control panel.

QuSAGE is the pathway analysis algorithm that has been designed to specifically address the issue of inter-gene correlation in complex contrasts. The majority of currently existing algorithms either explicitly assume independence of gene regulation or require a large number of replicates to adequately account for inter-gene correlations by performing sample permutations. For a detailed description of the algorithm and tuning parameters see the QuSAGE vignette

As per the QuSAGE vignette "Although QuSAGE does not require as many samples as permutation-based methods, the current implementation requires at least three degrees of freedom." In cases where the degrees of freedom falls below three (e.g. two replicas per condition), QuSAGE artificially sets the degrees of freedom to be three and issues a warning message that the results should be interpreted with caution because the P-values will be inflated, and may not accurately reflect the underlying data.

Two type of plots are generated by rnaseqDRaMA for each analyzed contrast -- QuSAGE Pathways Analysis and Pathway Genes plot.

14.1 QuSAGE Pathways Analysis Plot

QuSAGE Pathways Analysis plot shows differentially active pathway in each experimental contrast that can be selected in the Select a Contrast dropdown menu. This plots depicts changes in pathway-wide expression (pathway activity) sorted by the associated metric of statistical "significance" (P-value or FDR) -- Figure 14.1 or the effect size (pathway log fold change) -- Figure 14.2.

The size of a circle is proportional to the number of genes in the gene set. The color of a circle is proportional to the negative log10 value of "significance" metric. Hovering over the circle will display a tool tip with the MSigDB ID for a gene set, its size, significance metric, and pathway log fold change for the selected contrast. Finally, clicking on a circle will display the gene level changes for a desired set in the Pathway Genes plot below (Figure 14.3)).

QuSAGE Pathway Analysis plot. 
Gene sets were sorted by FDR

Figure 14.1: QuSAGE Pathway Analysis plot. Gene sets were sorted by FDR

QuSAGE Pathway Analysis plot. 
Gene sets were sorted by LogFC

Figure 14.2: QuSAGE Pathway Analysis plot. Gene sets were sorted by LogFC

The sorting method can be selected in the Select a Sorting Method dropdown menu in the Control Panel. Similar to other sections of the app, Pvalue or FDR? and FDR/Pvalue Threshold allow the selection of a significance metric and a cutoff value.

Finally, setting Display Significant Results Only? dropdown menu to TRUE filters out gene sets with the selected significance metric (P-value or FDR) higher than the cutoff values.

14.2 Pathway Genes plot

This plot shows gene level changes within a selected pathway. Genes are plotted as logFC +/- SD and ordered by logFC.

QuSAGE Pathway Genes plot. Genes sorted by LogFC

Figure 14.3: QuSAGE Pathway Genes plot. Genes sorted by LogFC

Pathways can be selected in the Search Pathways menu. This menu is also searchable. To search for pathways, click inside the Search Pathways dropdown box, delete the name of the current pathway using "Backspace" key and type the term of interest to initiate the search.

The plots can be saved by pressing Download plot (PDF) button. Summary table for all pathways or only for all genes within a selected pathways can be saved by pressing Download Selected Data. Finally, gene names for a gene set shown on the Pathway Genes plot can be copied to an internal clipboard using Copy selected gene button.