Do I need to redo PCA and UMAP if published cell type annotations are available?

1

Hello there,

I am working with a scRNA-seq dataset where the original publication already provides curated cell type annotations (in 3 levels in fact) in the metadata. My goal is to perform cell–cell communication analysis using CellChat.

Since (to my limited knowledge) CellChat only requires a normalized expression matrix and cell type labels, I am wondering whether it is necessary to redo dimensionality reduction and visualization steps such as PCA, clustering, and UMAP and … ?

or can I ,after normalization, directly use the existing cell type annotations for CellChat?

2

Hi @artintz

This is probably a better question for the people working in this domain everyday, and possibly the even the tool authors! A good place to start is here. → GitHub - jinworks/CellChat: R toolkit for inference, visualization and analysis of cell-cell communication from single-cell and spatially resolved transcriptomics

Then, once back in Galaxy, the tools will work the same as they do when working outside of Galaxy.

Others are still welcome to add in more! :slight_smile:

3

Hi @artintz what type of object does the publication provide? If it was already analyzed dataset, the object should have normalized expression, embeddings for PCA, UMAP, celltype annotation.

Best,

Pavan