Using infercnv R package

Question

[j.jacob1]

Hello,

I am a newbie to bioinformatics with limited programming experience.

I wish to use infercnv to separate out cancer (Medulloblastoma) cells co-cultured with wild-type cerebellar organoids in a scRNA-seq experiment with thousands of cells. I have the following conditions and controls:

1. wild-type cerebellar organoid (generated from iPSC differentiation protocol)
2. medulloblastoma cell line grown as a monolayer
3. medulloblastoma cell line grown as a spheroid
4. medulloblastoma cells co-cultured with wild-type cerebellar organoid - it is for this condition I would like to separate cancer cells from the wild-type cells.

The wiki for infercnv (https://github.com/broadinstitute/inferCNV/wiki/File-Definitions) mentions several inputs:

raw_counts_matrix which I can generate from the Seurat object. Would that be for condition 4?

annotations_file There is an example annotations file on the above website, but it is not clear to me how to generate this file. This question was asked by another individual on the infercnv GitHub (https://github.com/broadinstitute/infercnv/issues/215), but the answer provided ("The annotations_file is a basic table that is used to group together cells of the same type and patient/individual if you have that information. The format is described at https://github.com/broadinstitute/inferCNV/wiki/File-Definitions#sample-annotation-file . You can define cell types using marker gene analysis or another method of your choice, and define individuals from your metadata.") did not provide explanation for a newbie into how that table is generated. The simplest option would be to use the data from the Seurat object - reading further, an answer on the infercnv GitHub page suggests using the active.ident field to generate the annotation table, but doesn't mention how. The active.ident field has cell IDs and their cluster allocation.

For the gene_order_file input there is at least one provided by the Broad Institute who wrote infercnv.

ref_group_names is the last input. To keep things simple I wish to assign all the WT cerebellar organoid cells as 'normal'. In the 'annotations_file' do I simply populate the 2nd column with 'normal' for all the WT cell IDs? Obviously, I can only do that for the WT cerebellar organoid, but is that the right matrix to use for the 'raw_counts_matrix' input?

Thanks in advance.

[ccbb7aab4]

Hi Ram, I know this was posted six months ago. Did you have any trouble getting past step 08 whilst using R at all?

[Ram]

1. I did not create this post, you should be addressing j.jacob1
2. This post was created 12 months ago, not 6 months ago.
3. Don't add an answer unless you're actually answering the top level question. Use Add Comment or Add Reply instead as appropriate. I've moved your post to the right location now, but please be more careful in the future.