This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Integrate four scRNA-seq datas

What do I do with this warning when I'm putting together the data?

immune.anchors <- FindIntegrationAnchors(object.list = list(HLJ1, HLJ2,HLJ3,HLJ4), dims = 1:20)

Warning message:
In CheckDuplicateCellNames(object.list = object.list) :
  Some cell names are duplicated across objects provided. Renaming to enforce unique cell names.
rna-seq
• 2,506 views • link
updated by Biostar Biostar • written by 2962873004 • 2962873004

2 answers

You can ignore that. It's just saying there is some overlap in cell barcodes between samples, so it will add a suffix to the cell barcodes.

• 0 views • link
written by rpolicastro rpolicastro

But there's always a warning message after this step

• 0 views • link
written by 2962873004 • 2962873004

That's fine, there will almost always be matching cell barcodes since they are not completely randomly generated when 10X are synthesizing their beads.

• 0 views • link
written by rpolicastro rpolicastro

Is this warning okay

> immune.combined <- IntegrateData(anchorset = immune.anchors, dims = 1:20)
Warning: Adding a command log without an assay associated with it
• 0 views • link
updated by Ram Ram • written by 2962873004 • 2962873004

Yea, that can be ignored.

• 0 views • link
written by rpolicastro rpolicastro

Basically, when you have different experiments (or samples), some barcodes (cell names) in each experiment collide when merging them. So Seurat renames some cells to maintain unique names. So this can be ignored.

Regarding the other warning, seems like your initial objects were not normalized. While it can also be ignored, but better you normalize each of your HLJ1... before merging/integrating them. Make sure you have set the correct active assays for each object etc. (basically following the tutorial as it is)

https://satijalab.org/seurat/v3.0/integration.html

There is a step before integration, where each object is normalized.

for (i in 1:length(pancreas.list)) {
    pancreas.list[[i]] <- NormalizeData(pancreas.list[[i]], verbose = FALSE)
    pancreas.list[[i]] <- FindVariableFeatures(pancreas.list[[i]], selection.method = "vst", 
        nfeatures = 2000, verbose = FALSE)
}
• 0 views • link
written by GouthamAtla GouthamAtla

yeah。I have done the corresponding QC for each sample before I reintegrate it

• 0 views • link
written by 2962873004 • 2962873004

Thats not a QC, that's normalization.

for (i in 1:length(pancreas.list)) {

pancreas.list[[i]] <- NormalizeData(pancreas.list[[i]], verbose = FALSE)

pancreas.list[[i]] <- FindVariableFeatures(pancreas.list[[i]], selection.method = "vst", nfeatures = 2000, verbose = FALSE)

}

• 0 views • link
written by GouthamAtla GouthamAtla

Log in to answer this question.

More posts like this