cell ranger agrr
1
0
Entering edit mode
11 months ago

hello i have a question , i integrated two data ( 2 samples of B cells) with cell ranger aggr and with filtered matrix data from cell ranger aggr , i apply the seurat script but i dont know how to distingue between my two samples after clustering i can't use split.by because there is no row to use

R • 678 views
0
Entering edit mode

Are the two samples two different conditions/lines/...? If so, you don't need to run aggregate from cellranger. You can just go through the normal seurat integration workflow.

0
Entering edit mode

the samples are B cells, first condition is a subpopulation of the second condition. so i have to campare them to see what distingue this subpopulation from the other b cells, if i use the seurat integration workflow, the clustering will be the same for both

0
Entering edit mode

Regardless of the clustering, the sample identity will be stored as a meta-data variable. You can then use this with the FindMarkers (or related functions) to find transcriptional differences.

1
Entering edit mode
11 months ago

What software are you trying to use? Either combine with cellranger, or seurat, you don't use both. The Loupe output will know what library each barcode came from.

The barcodes of the aggregated data should all have a number appended to them to indicate which library they came from.

zcat barcodes.tsv.gz | tail
TTTCCTCCAAAGAATC-11
TTTGCGCAGCTAACTC-11
TTTGCGCGTACCGGCT-11
TTTGGTTAGCAATATG-11
TTTGGTTCACCGCTAG-11
TTTGGTTCAGGTCTCG-11
TTTGGTTTCCACTGGG-11
TTTGTCAGTACGCACC-11
TTTGTCATCAAGCCTA-11
TTTGTCATCACGAAGG-11


When you integrate with Seurat, you tell it what project name to assign to each sample, and that gets stored in the metadata when you integrate under "orig.ident"

>tail(combined.all[[]])
orig.ident nCount_RNA nFeature_RNA integrated_snn_res.0.2 seurat_clusters
TTTGGTTAGCAATATG-1_5  Untreated       3600         1157                      2               2
TTTGGTTCACCGCTAG-1_5  Untreated       3764         1198                      2               2
TTTGGTTCAGGTCTCG-1_5  Untreated       1941          679                      0               0
TTTGTCAGTACGCACC-1_5  Untreated       5478         1327                      2               2
TTTGTCATCACGAAGG-1_5  Untreated       2237          895                      4               4

0
Entering edit mode

exactly i would like to add this information in seurat , to tell him baracode with tail-1 are for exemple sample 1 and the other sample 2 , so i can after see which cluser belong to which sample but i cant find how to do it with seurat

0
Entering edit mode

so i use the output from cell ranger count

SSA.data <- Read10X(data.dir = "F:/tutorial seurat/cell count aggr/AGGSSA/outs/filtered_feature_bc_matrix") SSA <- CreateSeuratObject(counts = SSA.data, project = "SSA", min.cells = 3) dim(SSA) ... SSA <- RunUMAP(SSA, dims =1:15) DimPlot(SSA, reduction = "umap", group.by = "orig.ident")

my orig.ident are all SSA so r can't use this function

           orig.ident nCount_RNA    nFeature_RNA    percent.mt  RNA_snn_res.0.8 seurat_clusters AAACCTGAGGCTCATT-1  SSA     1404     779         9.97150997150997       1                               1

0
Entering edit mode

Well, sure, if you read in the data from aggr and assign the project only once, it will all have one project name. You can figure out from the barcode suffixes which barcodes came from which samples, add your own column to the meta data.