search

neurorestore / Libra

MIT License
153 stars 25 forks source link

Replication, Label, and Cell Type for DEG #27

Open Sayyam-Shah opened 2 years ago

Sayyam-Shah commented 2 years ago

Hello @jordansquair,

Thank you for the amazing tool!

I have 6 samples. 3 are malignant and 3 are non-malignant. I subsetted certain cell types and I want to run DEG between the malignant and non-malignant samples (malignant vs. non-malignant). May you please explain to me what the difference is between biological replicates and sample labels. All the cells in the subset are B cells.

Also, I am working with scanpy so would I input my obs data frame? Can I input the expression matrix as a data frame as well?

Best Regards, Sayyam

jordansquair commented 2 years ago

replicate_col is your biological replicate (sample ID) and label_col is your group.

You can input the expression matrix with the accompanying meta data, or convert to a Seurat object.

Sayyam-Shah commented 2 years ago

Hello, that makes sense. I'll convert it to a Seurat object with SeuratDisk.

May you please clarify what you mean by group? In my case would that be malignant vs non-malignant. My cell type would be the same for all cells. As a result, will it be performing DEG across label_col?

jordansquair commented 2 years ago

group would be malignant vs. non-maligant. set cell_type = 'B cells' and label to whatever meta data vector has the malignant status. It will run DEG across the label (maligancy)

Sayyam-Shah commented 2 years ago

Thank you so much!