This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Classify cells using seurat

Hi,

I am working on single cell data and I have a general question. Monocle has a function 'classifyCells' to assign celltype using known marker genes. Here is the bit from the documentation

cth <- newCellTypeHierarchy()

MYF5_id <- row.names(subset(fData(cds), gene_short_name == "MYF5"))
ANPEP_id <- row.names(subset(fData(cds), gene_short_name == "ANPEP"))

cth <- addCellType(cth, "Myoblast", classify_func =
    function(x) { x[MYF5_id,] >= 1 })
cth <- addCellType(cth, "Fibroblast", classify_func =
    function(x) { x[MYF5_id,] < 1 & x[ANPEP_id,] > 1 } )

cds <- classifyCells(cds, cth, 0.1)

Is there a similar function is Seurat. Something that allows to classify cells by cell type and adds it to the meta-data in the seurat object?

scrna seurat monocle
• 5,161 views • link
updated by dppb05 • dppb05 • written by AB AB

I have used SetIdent() to do something like that.

• 0 views • link
written by Madelaine Gogol Madelaine Gogol

1 answer

Is there a similar function is Seurat. Something that allows to classify cells by cell type and adds it to the meta-data in the seurat object?

There is ClassifyCells function. But keep in mind that this function works differently from monocle::classifyCells. Monocle's docs, as you have probably read, give a good high-level explanation on their classification approach.

For Seurat::ClassifyCells you need some labelled data that will be used to train Random Forest classifiers. Those trained classifiers will then be used to classify your unlabelled data. You could subset your Seurat object (using SubsetData) based on some marker genes and set the ident (using SetIdent) of this subset according to those markers, then use that as your training set.

• 1 views • link
written by dppb05 • dppb05

Log in to answer this question.

More posts like this