Mapping failed because of application errors

How To MapMyCells
1

Hi! I am really excited to use this tool but have been encountering issues at the last stage of the analysis and receiving a “Mapping Failed. Mapping algorithm failed because of application errors.”
Here is the requested info:

  • Operating system and web browser: Windows 11 and Chrome
  • Run ID: 1697776005895-695f859a-5df9-4a1e-8fab-97eb3aab835e
  • Information on the data: This is mouse snRNAseq data generated using a combinatorial indexing approach (sciRNAseq). The minimal h5ad was generated in R with just the cell-by-gene table. Column names are ENSEMBL IDs (e.g. ENSMUSG00000051951) and row names are cell barcodes. The AnnData object n_obs x n_vars = 50000 x 22004 with a total size of 327.4MB.
  • Instructions on how to reproduce the issue: I uploaded the h5ad to the site and am using the 10x Whole Mouse Brain as the reference taxonomy and hierarchical mapping as the mapping algorithm. The job proceeds through ‘input file validation’ and ‘algorithm run’, but errors during the ‘Mapping completed’ step.
    Any suggestions would be greatly appreciated! Thanks.
2

Hi,

It looks like your h5ad file contains negative gene expression values (minimum value is -7). This is a problem for our algorithm as we assume that the inputs are raw counts which we then convert to log2(counts per million + 1). Obviously, the log2 is not going to handle -7 very well.

Hope this isn’t an absolute show stopper for you.

Cheers,

Scott Daniel

3

Thanks for catching that. I fixed it and got it work, thanks for your help!

4

I am also very excited about using this tool but seem to be experiencing a similar issue. The Run ID is 1731885286215-b912ca5e-bfd3-475d-9c70-eba3fddf5fa2. Does anyone have any advice/suggestions? Thanks in advance for any help!

5

@Sumira83

Your job failed for a different reason. I am going to open a new thread where we can discuss it specifically. This should make it easier for people with errors like yours to find our conversation. I will post a link to the new conversation here shortly.

6

@Sumira83

Here is where I have moved the conversation surrounding the error you encountered.

7

Hi! I am experiencing some issues too, I am not sure what I have to check. Any advice or guidance is highly appreciated! Thanks!
This is the Run ID: 1731966422975-f370b2ca-fb7c-49ca-8b2b-dc4712a238bd

8

@RebecaO

You have the same error that Sumira83 has. Your h5ad file is missing a metadata field that, according to the anndata documentation, is supposed to be present in all h5ad files.

What software/tool did you use to create your h5ad file?

9

For anyone else who finds this thread, I just want to call out this post, which shows how to download the failure log from a failed MapMyCells job. That wouldn’t have helped much in the cases above (the error is still pretty in-the-weeds), but for more anodyne errors, it might help get users unstuck without having to wait for a back-and-forth on this forum.

10

I am working on R, but I’ll work on that! Thanks for letting me know how to download the failure log.

11

@RebecaO
If I ask “what library in R did you use to create the .h5ad file?” is there a succinct answer? Again: just curious

12

Hi Daniel, the function in the code I used is writeH5AD() from the zellkonverter package.