invalid class "CellDataSet" object: sampleNames differ between assayData and phenoData

[sylestiel]

Hi Tim,

I was trying to generate a cds object with a view to getting the co-accessible sites. I'm getting the following error:

int.cds <- as.CellDataSet(integrated,assay="ATAC", reduction = "wnn.umap")

Error in validObject(.Object) :
invalid class "CellDataSet" object: sampleNames differ between assayData and phenoData
In addition: Warning messages:
1: In .sparse2dense(e1) :
sparse->dense coercion: allocating vector of size 24.5 GiB
2: In .sparse2dense(e2) :
sparse->dense coercion: allocating vector of size 24.5 GiB

What follows are the specifics of my object:

integrated

An object of class Seurat
218108 features across 20983 samples within 4 assays
Active assay: ATAC (156711 features, 156711 variable features)
2 layers present: counts, data
3 other assays present: RNA, SCT, integrated

5 dimensional reductions calculated: pca, umap.rna, lsi, umap.atac, wnn.umap

> integrated@reductions

$pca
A dimensional reduction object with key PC_
Number of dimensions: 50
Number of cells: 20983
Projected dimensional reduction calculated: FALSE
Jackstraw run: FALSE
Computed using assay: integrated

$umap.rna
A dimensional reduction object with key rnaUMAP_
Number of dimensions: 2
Number of cells: 20983
Projected dimensional reduction calculated: FALSE
Jackstraw run: FALSE
Computed using assay: integrated

$lsi
A dimensional reduction object with key LSI_
Number of dimensions: 50
Number of cells: 20983
Projected dimensional reduction calculated: FALSE
Jackstraw run: FALSE
Computed using assay: ATAC

$umap.atac
A dimensional reduction object with key atacUMAP_
Number of dimensions: 2
Number of cells: 20983
Projected dimensional reduction calculated: FALSE
Jackstraw run: FALSE
Computed using assay: ATAC

$wnn.umap
A dimensional reduction object with key wnnUMAP_
Number of dimensions: 2
Number of cells: 20983
Projected dimensional reduction calculated: FALSE
Jackstraw run: FALSE
Computed using assay: ATAC

I integrated the RNA modality from 3 multiome assays but NOT the ATAC modalities since these were multiome datasets.. So there is overlap when I look at the wnn.umap projection and umap.rna projection but NOT with the umap.atac projection.
Is that to be expected? Also to run LinkPeaks() I do have to establish the conns correct?
So which reduction should I be using? wnn.umap or umap.atac.

What is the error:
invalid class "CellDataSet" object: sampleNames differ between assayData and phenoData referring to? How may I fix it?

Despite adding the chr prefix to the ranges in the ATAC component of the object I'm still unable to run the Motif analysis. As a reminder I'm using the Macaca mulatta datasets and used the gtf file at Ensembl for Mmul_10 for annotation purposes.

Thanks for your time!

[timoast]

You need to use the as.cell_data_set function instead

Also to run LinkPeaks() I do have to establish the conns correct?

No, if you want to run LinkPeaks() there's no need to run Cicero

[sylestiel]

Getting this error
integrated <- LinkPeaks(

• object = integrated,
• peak.assay = "ATAC",
• expression.assay = "SCT",
• genes.use = c("LEF1")
• )
  Testing 1 genes and 156706 peaks
  | | 0 % ~calculating Error in density.default(x = query.feature[[featmatch]], kernel = "gaussian", :
  argument 'x' must be numeric

convert to CellDataSet format and make the cicero object

int.cds <- as.cell_data_set(x = integrated)
Error in as.cell_data_set(x = integrated) :
could not find function "as.cell_data_set"

convert to CellDataSet format and make the cicero object

int.cds <- as.CellDataSet(x = integrated)
Error in validObject(.Object) :
invalid class "CellDataSet" object: sampleNames differ between assayData and phenoData
In addition: Warning messages:
1: In .sparse2dense(e1) :
sparse->dense coercion: allocating vector of size 24.5 GiB
2: In .sparse2dense(e2) :
sparse->dense coercion: allocating vector of size 24.5 GiB

[sylestiel]

Hi@ Tim,

Still awaiting for a way to troubleshoot these problems.

Thanks,

[timoast]

Hi @sylestiel, please open an issue containing the full code you're running and the output of sessionInfo()

[sylestiel]

Full code as in from the very start of analysis?