How to ID cell types using reference scRNAseq dataset

Question

[FantasticAI]

I'm very new to scRNAseq analysis, and I really appreciate any inputs and comments. Your feedback would help me a lot.

I have been given 3 groups data (4 normal, 8 disease and 8 treatment samples), what is the workflow to identify how man cell types and also ID the cell types using reference dataset(http://mergeomics.research.idre.ucla.edu/PVDSingleCell/CellBrowser/?ds=PAHRatLungs#) in each group?

This question might sounds too general. Specifically, I wonder whether I need to merge data within each group and then perform analysis for each group separately? or I should merge and integrate all 20 samples together? For QC, from my limited understanding, I should perform QC on each samples (20 in total) before doing any computation analysis, right?

Thank you so much for the kindly help!

[ATpoint]

Cross-posted: https://bioinformatics.stackexchange.com/questions/15847/whats-the-difference-between-seurat-merge-and-integration

Answer 1

[ATpoint]

Yes, QC should be done on each sample individually. OSCA, the Bioconductor single-cell compendium https://bioconductor.org/books/release/OSCA/ is a great read. Spend some quality time going through it, it should cover 99% of what you need.

[FantasticAI]

Thank you for the rely, appreciate it very much. I do have another question regarding to the tutorials. I noticed that there is another tool on scRNAseq, called "Seurat". I wonder which one is more easy to use, Seurat or Bioconductor? As far as I know they are similar on scRNAseq analysis.

[ATpoint]

I personally prefer Bioconductor over Seurat, but Seurat has a large userbase itself. The linked resource is outstandingly comprehensive with lots of background information on the "why's" in addition to the "how's", one major reason why I prefer it. Seurat is also fine though.

[FantasticAI]

Thank you so much