Meta-analysis: inter-file comparison¶

Meta-analysis allows you to create a set of plots to help compare the same cell population across files.

---

Set-up Meta-analysis¶

After you have finished analyzing all your files, you can compare the same cell populations across your files by initiating a Meta-analysis. On the Experiments page, Workflow tab :

1. Click on "Create meta-analysis".
   • Section "Select analyses":
     • On the pop-up, select the files you want to compare by clicking on the corresponding checkboxes.
     • Click "Next".
   • Section "Select groups":
     • Your files are automatically categorized into a group called "Group 1".
     • If your files can be categorized into more than one group (e.g. healthy, sick):
       • Click "+ Add" on the top right corner of the file table to add more groups.
       • You can rename the groups by clicking on the icon or delete the group by clicking on the icon.
     • Once you are done, click "Next".
   • Section "Create meta-analysis":
     • Enter the "Name" or the identifier of your meta-analysis for future reference.
     • Enter the "Description" for your meta-analysis.
     • For "Dimensionality reduction", select the method you want to use (UMAP or t-SNE).
     • Select "Batch effect removal" if you want to remove batch effects between files before combining them together for the dimensionality reduction visualization. Batch effect removal is NOT applied for any other visualization nor is it applied for calculating statistics. This is to avoid inducing human artifacts into your results.
   • Click "Create".

---

Define Meta-analysis plots¶

From your Experiments page , click on the "Meta-analysis" tab to view the meta-analysis you have created.

1. Click on the name of the Meta-anlaysis you want to view.
2. On the Meta-analysis plots page , you can create a new plot by clicking on the icon.
   • You can select the type of plot you want to create from the first dropdown menu.
   • After the plot type has been selected, plot parameters will appear below for you to adjust.
   • After adjusting parameters, click "Create" to finish.
   • If you want to edit your plots, click the icon on the top right corner of your plot.

The types of plots available in Meta-analysis include:

• "Clusters frequency by sample": A stacked barplot with your files on the Y axis, and your cell populations' frequency on the X axis.
• "Dimensionality reduction": The 2D UMAP or t-SNE plot containing cells from all of your files.
• "Box plot": A boxplot with your files on the X axis and your cell populations' FI on the X axis.
• "Frequency chart": A barplot with your cell populations' mean or median FI on the Y axis and frequency on the Y axis.
• "Spider plot": A spiderplot describing your cell populations' mean or median FI over your Files.
• "Volcano plot": A scatterplot with the logged p-value on the Y axis and logged fold change on the X axis to compare abundance of cell populations between selected file groups.