Could you please tell me how should I analyze the difference between KO and WT mouse sequenced by 10X genomics?

Question

[wayj86]

Dear all,

10X Single-cell RNA-seq were performed on four mouse samples, two KOs (only one gene was knocked out) and two WTs. KO1 and KO2 have the same weeks of age with WT1 and WT2, respectively. The purpose of this study is to investigate the difference between KO and WT, as well as the impact of Knocking-out on the expression of related genes.

The sequencing were performed via the workflow "Multiple Samples, Multiple GEM Well, One Flowcell". I ran Cellranger count four times separately for the four samples, then I don't know how to do for the next step.

---

Option 1. Should I use Cellranger aggr to aggregate KO1 and KO2 to a single KO output, and then WT1 and WT2 to a single WT output?

Option 2. Or should I directly use Cellranger aggr to aggregate the four samples to a single output, containing all KOs and WTs?

Option 3. Or should I use Cellranger aggr to aggregate KO1 and WT1, as well as KO2 and WT2, because KO1 and WT1 were sequenced in a same flowcell?

---

After Cellranger aggr (no matter option 1, 2 or 3), how should I perform the next differential analysis? Could Seurat do that?

Thanks a million for your attention,

Stanley

Answer 1

[Kristoffer Vitting-Seerup]

I highly recomend Seurat to do the analysis downstream of quantification - it is easy to use, fast, have a very detailed workflows one can follow (fx this) and most importantly it is one of the best tools for clustering single cell data (according to this article).

With regards to how to run CellRanger I think Seurat needs 1 set of files ( see this documentation page) but I could be wrong.

[suvratha]

can Cell Ranger do differential expression in its pipeline or do you have to use Seurat for doing differential expression analysis?