Normalization of index sort data in Seurat

Question

[f.morgana]

Hi there,

I am new in the field of bioinformatics and R and have been trying to do the multi-modal vignette from the Satija lab (Seurat) but without success. See link to the vignette here: https://satijalab.org/seurat/v3.1/multimodal_vignette.html

I have performed scRNAseq in combination with index sorting. Now I am trying to link the index sort data with the scRNAseq data. However, I have not been successful in doing so with the vignette. I cannot seem to get the CreateAssayObject function to work, as I get "Error: Cannot add a different number of cells than already present".

Furthermore, I do not fully understand all the steps.

I understand that they first of course load the data. However, what do they do in the second line ("rownames(x=pbmc10k.data[["Antibody Capture"]]...")? Is "Antibody Capture" already a column in their metadata or are they adding it here?

pbmc10k.data <- Read10X(data.dir = "../data/pbmc10k/filtered_feature_bc_matrix/") rownames(x = pbmc10k.data[["Antibody Capture"]]) <- gsub(pattern = "_[control_]*TotalSeqB", replacement = "", x = rownames(x = pbmc10k.data[["Antibody Capture"]]))

I also understand that after that they create their object and subsequently normalize their data:

pbmc10k <- CreateSeuratObject(counts = pbmc10k.data[["Gene Expression"]], min.cells = 3, min.features = 200) pbmc10k <- NormalizeData(pbmc10k)

However, what are they doing as they add the ADT data?

pbmc10k[["ADT"]] <- CreateAssayObject(pbmc10k.data[["Antibody Capture"]][, colnames(x = pbmc10k)]) pbmc10k <- NormalizeData(pbmc10k, assay = "ADT", normalization.method = "CLR")

Similar question here, are they here adding a column called "ADT" to their metadata? Also, which normalization method to use in the case of index sort data as opposed to ADT data?

Looking forward to any suggestions!

Many thanks in advance, Florencia

[f.morgana]

Dear Igor,

Many thanks for your answer, after some additional reading and thanks to your answer it became much clearer to me and I managed to apply the tutorial to my data.

Best wishes, Florencia

Answer 1

[igor]

Is "Antibody Capture" already a column in their metadata or are they adding it here?

You should download the example data they provide to better understand it. That way you don't have to guess what is happening.

You should also read the function description. For example, Read10X() returns a list containing a sparse matrix of the data from each type will be returned. Basically, you get 2 matrices, one for "Gene Expression" and one for "Antibody Capture". Those labels are based on the labels that Cell Ranger assigned.

However, what are they doing as they add the ADT data?

They are calling CreateAssayObject() and giving it the "Antibody Capture" matrix.

Similar question here, are they here adding a column called "ADT" to their metadata?

"ADT" is an "assay" within the Seurat object, not a column. Basically, assays are used as a way to keep RNA and ADT data separate but still have it associated with the same cells.