Mapping failed because of application errors troubleshooting

Operating system and browser: Windows 11, chrome
Run ID: 1698255812324-4d53ffc5-9c7c-4dff-b0b4-e4caf0923569
Information on the data: Mouse single nuclei sequencing data converted from a Seurat object originally in R. The data set only contains neurons which were parsed out based on marker genes. Column names are cells and rows are mouse genes. It has some meta data on the clustering so maybe that is a problem?
Instructions on how to reproduce your issue: The .h5ad file gets uploaded but the error happens when i try the hierarchical mapping during the algorithm run step.

Thank you for your help in advance!

@suhjinlee , thank you for trying MapMyCells.

MapMyCells requires that for the input file, rows are cells and columns are genes.

From your description, it sounds like the opposite is the case for your file. This would make sense since that’d be the default for Seurat/R.

Can you please transpose your matrix to the expected format and try again?

Let us know if this doesn’t fix your issues and we’re happy to continue troubleshooting.


1 Like

I actually think that the “columns are cells, rows are genes” is a typo. I’ve looked at the input file (neurons.h5ad, right?) and the columns and rows look right.

The actual problem (or another problem) is that the minimum expression value is -2.3. Our algorithm assumes inputs are raw counts which get converted to Counts Per Million and then scaled to log2(CPM+1). The log2 means that the expression matrix cannot have negative values.

Hope this helps,

Scott Daniel

1 Like

Hello. I changed the default assay to be “RNA” and think now I am using raw count numbers.

I unfortunately encountered another error during the algorithm run process.

This time the run id is : 1698270854411-45be3401-62f8-4069-981e-bb852b4e0036

I used a new file (also named neurons.h5ad) and checked that the rows are samples and columns are genes. Would you be able to help troubleshoot this?

Thank you so much!

It looks like you still have negative values in your expression matrix. Can you open the file on your local compute and check the minimum value of the ‘X’ matrix in the h5ad file?