How to acquire anchor features for multiple dataset single-celll analysis in Seurat 3.0?
0
0
Entering edit mode
2.3 years ago

I am attempting to analyze sequencing from mouse and chicken samples in Seurat by following the multi-dataset integration vignette on the Seurat website: https://satijalab.org/seurat/v3.1/integration.html. In their example they use FindIntegrationAnchors(object.list = reference.list, dims = 1:30) to find anchors for integrating human pancreatic samples from four different platforms. This results in many anchor features that are used in subsequent steps. However, I am attempting to recreate the analysis with mouse and chicken samples. In the FindIntegrationAnchors step, only three anchors are returned, rendering the following steps unworkable. I assume given the greater disparity between mouse and chicken expression than between different human samples, that there will be fewer anchors, but is there a way to collect more? Seurat claims this workflow can work for different species. Has anyone had success with such an analysis?

RNA-Seq rna-seq genome sequencing next-gen • 1.9k views
ADD COMMENT
0
Entering edit mode

Are your samples comparable in terms of the cells being assayed? Are they all like lung or blood or bone marrow... or of completely different origin from these two animals? Please give some details and show the code you used.

ADD REPLY
0
Entering edit mode

They are both embryonic limb cells. My code follows the vignette. The datasets are quite large.

library(Seurat)

mouse.data <- Read10X(data.dir = "/media/david/mousecount/outs/filtered_feature_bc_matrix")
chicken.data <- Read10X(data.dir = "/media/david/chickcount/outs/filtered_feature_bc_matrix")
mouse = CreateSeuratObject(counts = mouse.data)
chicken = CreateSeuratObject(counts = chicken.data)

species.list <- c(mouse, chicken)

for (i in 1:length(species.list)) {
  species.list[[i]] <- NormalizeData(species.list[[i]], verbose = FALSE)
  species.list[[i]] <- FindVariableFeatures(species.list[[i]], selection.method = "vst", 
                                             nfeatures = 2000, verbose = FALSE)
}

SelectIntegrationFeatures(object.list = species.list, verbose = TRUE)

reference.list <- species.list
species.anchors <- FindIntegrationAnchors(object.list = reference.list, assay=c("RNA", "RNA"), dims = 1:30)
ADD REPLY
1
Entering edit mode

How many genes are common between the two Seurat objects?

ADD REPLY
0
Entering edit mode

Ahh there are only 13. I suppose that is an issue.

Is there a tool I can use to discover the orthologs between the two species and then reannotate the genes?

ADD REPLY
1
Entering edit mode

That it is. You will have to do some gene ID conversions to get more overlap between the two.

ADD REPLY
0
Entering edit mode

Sorry just edited my response above before reading,

Is there a tool I can use to discover the orthologs between the two species and then reannotate the genes?

ADD REPLY
0
Entering edit mode

biomart is probably the easiest way. There are several other questions regarding how to do this on Biostars as well if you do a little searching.

ADD REPLY

Login before adding your answer.

Traffic: 1237 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6