Skip to content

Reduction of dimensions

Now, you have a clean expression matrix and you want to identify cluster of populations. First, you'll need to explain as much variability in as few dimensions as possible. Because, at the moment, you can only describe your cells in 2000 dimensions (maybe a little complex for the poor human eye and also for a computer!). To learn more about different methods, you must read the following page: Reduction of dimensionality.

Do it yourself!

Render your RMD/QMD

To complete this week you'll need to :

  • 1. Compute the PCA
  • 2. Perform the JackStraw analysis
  • 3. Utilize JackStraw and Elbow visualization to choose the number of PCs to retain explain your choice
  • 4. Compute the UMAP and perform different visualizations with DimPlot and FeaturePlot, have fun!

IMPORTANT

Please note you must be explicative in your cutoff choices and provide detailed explanations for each step of your thought process In general, try to explain in your own words for each step of your analysis!

Add your RMD/QMD in your Trello card.

Thank you for your attention and see you next week 👏 👏 👏