18 DIY Lab

18.1 DIY Lab

Hopefully now you have a “feel” for what scRNA-seq analysis entails. Today we will work in groups to analyze a publicly available data set: IDH-mutated gliomas. Your data includes: IDH_A_processed_data_portal.txt - the TPM matrix, already in log scale IDH_A_cell_type_assignment_portal.txt - a classification of cells to malignant and non-malignant groups, and to tumors You also have files for different signatures.

What would you like to focus on today? If it’s clustering and identifying cell populations we reccomend you use all the data and try to distinguish the different cell types. If it’s combining a few datasets and work on batch correction, we reccomend you focus on all malignant cells and work on the batch effects among tumors and identifying cell populations shared among them. If it’s identifying subtle differenced in clustering and trying also to identify lineages we reccomend you use one tumor. Most importantly, feel free to explore what ever interest you!

Create a seurat object filtering out the very extreme cases.

## Warning: Feature names cannot have underscores ('_'), replacing with dashes
## ('-')

You can start by filtering extreme outliers if you would like (just replace the ‘?’)

And normalize:

NOTICE: THESE FOLLOWING COMMANDS WILL ONLY WORK AFTER CREATEING UMAP/tSNE Here are useful markers that can be used to explore the clusters after you will cluster the data

And a few useful commands, that can be used after clustering:

If you want to focus on just malignant cells, or just one tumor at a time

Ok, you are ready! Discuss what you wish to explore and go for it!