Volcano plot - stjude/proteinpaint GitHub Wiki

The volcano plot visualizes differential analysis results by plotting statistical significance (-log10(p-value)) against fold change (log2(fc)).

How this Plot Works

The colors represent the selected control and case group colors as defined in the Groups tab of the portal UI. If no selections were made, groups default to red and black.

Data points with zero p-values are plotted at the lowest nonzero p-value in the plot data.

How to Launch

Follow the instructions from the differential analysis app wiki page.

Interactive Features

Users can highlight genes, add confounders, launch the box plot, etc. using the action buttons, p value table, and configure the display from the burger menu to the left of the plot.

Using the P-Value Table

Hovering over the rows in the p value highlights the data point in the plot. Clicking on the row allows the highlight to persist.

For gene expression, the table shows Gene Name, fold-change, and p-values. For DNA methylation, the table shows separate Promoter and Gene(s) columns, since the analysis is at the promoter level.

Launch Box Plot

Clicking on a data point opens the data in both groups. This feature is available for gene expression only (e.g., clicking on a gene will display that gene in each group).

Action buttons

The buttons above the default plot display the images and statistics as well as show/hide the p value table to the left.

Gene Expression

For gene expression, two additional buttons appear: Confounding factors and Genes.

• Confounding factors: Add confounding factors by selecting a term from menu and click submit. The plot will rerender with data accounting for the confounding factors. Note, the button is not available when:
  • Wilcoxon is the selected method
  • edgR is the selected method and there are greater than 100 total samples.
• Genes: Highlight genes in the volcano plot by inputing in the search bar or use a list from pre-defined datasets. Create a gene list from the UI and click submit to see the data points highlighted in the volcano. If a gene selected is not available, the display will show the gene with a line through. The Cancel highlight button in the top right removes the highlight.

DNA Methylation

For DNA methylation, a Confounding factors button is available. Confounders (up to 2, continuous or discrete) can be added to adjust the limma design matrix for batch effects, age, sex, etc.

Controls

Click on the burger menu in the upper lefthand side to see the controls. The default plot allows users to change the colors, size, fold change, and p-value type (i.e. original or adjusted).

Gene Expression

For gene expression, additional controls to change the method, minimum read count, and total read count are available in the burger menu.

DNA Methylation

For DNA methylation, a Min samples per group control is available. This sets the minimum number of non-NA samples required in each group for a promoter to be included in the analysis (default: 3). Limma requires at least 3 samples per group for reliable variance estimation.