Hello, I am following Seurat's pipeline https://satijalab.org/seurat/v3.0/immune_alignment.html for integrating cross-specie data.
In the code below, object.list inputs a list of Seurat objects.It gives warning:
skin.anchors <- FindIntegrationAnchors(object.list = list(skin2, skin3,neoskin))
Warning message: In RunCCA.Seurat(object1 = object.1, object2 = object.2, ... : Fewer than 50 features used as input for CCA. Requested number is larger than the number of available items (3). Setting to 3. X 50
Then following output by IntegrateData() function:
> skin.integrated <- IntegrateData(anchorset = skin.anchors, dims = 1:30)
Output:
Merging dataset 2 into 1 Extracting anchors for merged samples Finding integration vectors Warning in irlba(A = t(x = object), nv = npcs, ...) : You're computing too large a percentage of total singular values, use a standard svd instead. Finding integration vector weights Error in Embeddings(reduction)[nn.cells2, dims] : subscript out of bounds
Can anyone please help resolving the error? Thank You in advance!
Are the gene names the same across the three samples?
No, gene names are different across the datasets. How can be they exactly same, as the datasets are of different species? Also, anchors were calculated, although very less in number. What is expected here?
How can you expect them to find anchors if you don't use common gene names? Integration requires the comparison of common features. In this case, that'd be orthologous gene. So save your current gene names somewhere in the metadata, convert one species' orthologous genes to the other or vice versa, perform your integration, and then slap the old gene names back on.
Okay, I got the issue. Thank you so much for the provided direction, it really helped.
check which assay you are using for the integration. I had a similar error once, when I was integrating "integrated" objects without setting the assay to "SCT" (
DefaultAssay(object) <- "SCT").Hey, I guess what you are mentioning is for Seurat's SCTransform pipeline. In my case, initially RNA assay would be used for integration purpose which will create a new "corrected" matrix for integrated datasets ,used for comparative downstream analysis.