Skip to content

Not certain how MuSiC was run with Pbmc8k on real bulk #5

Description

@SergejRuff

Hello,

I recently read your DISSECT paper and had a question about how you ran MuSiC on the real PBMC8k dataset to generate Figure 2.

As far as I understand, MuSiC requires multiple subjects in the single-cell reference data. However, the PBMC8k dataset appears to contain cells from only one donor. I tried running MuSiC using the code from your repository:

> colnames(df_celltypes) <- c("cellType", "SubjectName")
> df_celltypes$cellTypeID <- df_celltypes$cellType
> df_celltypes$subjectID <- df_celltypes$SubjectName
> rownames(df_celltypes) <- colnames(df)
> 
> pheno_data <- AnnotatedDataFrame(data = df_celltypes)
> eset <- ExpressionSet(as.matrix(df), phenoData = pheno_data)
> 
> est.prop <- music_prop(
>   bulk.eset = eset_bulk,
>   sc.eset = eset,
>   clusters = "cellType",
>   select.ct = unique(df_celltypes$cellType),
>   verbose = FALSE
> )

In the repository code, all cells seem to be assigned to "donor1". When I try this, MuSiC crashes because the reference contains only a single donor.

Could you explain how you performed the MuSiC analysis for this dataset? I may be missing something, but I would like to understand how the single-donor issue was handled.

Best regards,
Sergej Ruff

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions