Trouble creating seurat obj from matrix

[JunlinJLiu]

I don't know if this has been asked before but I could find anything yet...

I tried to read an hdf5 sc-seq file into Seurat but it was too large so I wanted to load the first 10000 data. The data is in the following format: h5ls("expression_matrix.hdf5")

group    name       otype  dclass             dim
0     /    data   H5I_GROUP                        
1 /data  counts H5I_DATASET INTEGER 1169320 x 31053
2 /data    gene H5I_DATASET  STRING           31053
3 /data samples H5I_DATASET  STRING         1169320
4 /data   shape H5I_DATASET INTEGER               2

According to the structure I used the following code to load the first 10000 data:

counts <- h5read("expression_matrix.hdf5", "/data/counts", start = c(1,1), count = c(10000, 31053)) 
genes <- h5read("expression_matrix.hdf5", "/data/gene") # first 10,000 genes 
samples <- h5read("expression_matrix.hdf5", "/data/samples")  [1:10000]  
counts <- t(counts) 
rownames(counts) <- genes 
colnames(counts) <- samples 
Mouse <- CreateSeuratObject(counts = counts, project = "mouse_sample", min.cells = 3, min.features = 200)

The last step gave me this error:

Counts matrix provided is not sparse. Creating V5 assay in Seurat Object.
Error in slot(object = object, name = "features")[[layer]] <- features :  more elements supplied than there are to replace

I had no clue what caused this error. Can anybody give me a hint?

[samuel-marsh]

Hi,

Not member of dev team but hopefully can be helpful. Have you tried reading in file using Seurat V5 beta and the BPCells package (https://satijalab.org/seurat/articles/seurat5_bpcells_interaction_vignette)? Do you still have issues using that pipeline?

Best, Sam

[JunlinJLiu]

Have you tried reading in file using Seurat V5 beta and the BPCells package (https://satijalab.org/seurat/articles/seurat5_bpcells_interaction_vignette)? Do you still have issues using that pipeline?

Thanks for the reply Sam! I think the problem with my data is that everything is nested under data so instead of finding counts, gene, and sample we have to look at data/counts etc. And I couldn't find anything in these methods that let me adjust that path. And that's basically why I'm manually loading the matrix and converting them into seurat.

[clee700]

Did you find a solution? I have the same issue.

[clee700]

I still do not know what caused the issue or what it meant, but creating it as an Assay Object then turning that into a Seurat object seems to have worked.

dat = CreateAssayObject(gmat.use) dat = CreateSeuratObject(dat)

[erzakiev]

I encounter a similar error, while also reading in an h5 data in CSR/CSC format... The CreateAssayObject workaround by clee700 worked...

Error in slot(object = object, name = "cells")[[layer]] <- cells : 
  more elements supplied than there are to replace

[zhangfeiran]

It turns out this is due to row/colnames of counts must be character but not array, thus

genes <- as.character(genes)
cells <- as.character(cells)

before assigning the row/colnames solved the problem.

[mhkowalski]

Hello,

Thanks everyone for your contributions. I'm closing this now as it appears to be resolved, but please open a new issue if you are continuing to have trouble. Thanks!

[yojetsharma]

I have been facing this. The suggested solutions didn't resolve this issue for me.