The dataset used for the demonstration hagai_toy reported an error

[Taoo777]

@skinnider @jordansquair @AlanTeoYueYang @neurorestore
Thanks to this useful tool, I have been able to use Libra smoothly before. I reinstalled Libra today due to R environmental reasons, but found an Error when I ran Libra:

Error  in 'group_by()' :
! Must group by variables found in `.data`.
✖ Column `cell_type` is not found. 

To check if this is a problem with my dataset, I ran run_de: DE = run_de(hagai_toy) with the Libra built-in dataset hagai_toy but still got the same error, suggesting that there might be a problem with Libra itself. I noticed that Libra has been updated recently. Could you please check if there are any bugs?

[AlanTeoYueYang]

Hi, we have identified and fixed the bug.

[Taoo777]

@AlanTeoYueYang
Thank you very much for your quick reply! I am now able to use Libra normally.

As a small suggestion, I would like to know if it is possible to add gene expression level information, such as the baseMean of DESeq2 or others, to the final result of run_de. This is because I noticed that the most differentially expressed genes tend to have very low expression levels (not even visible on the UMAP plot), and in fact, we focus more on genes that have a certain expression level and are also differentially expressed. So I thought that if the final result could provide information about gene expression levels, it would help users filter out genes that are too low to be of interest to them. I would really appreciate it if you would consider this!

[Taoo777]

I noticed that the Libra:::pseudobulk_de function do not clean up the results of the differential expression analysis, so I tried using this function: DE_test = Libra:::pseudobulk_de(expr,meta = meta), but I got the following error:
Error in map():
ℹ In index: 1.
Caused by error in expr0 %*% mm:
! non-conformable arguments

However, when I use run_de: DE = run_de(expr,meta = meta) with the same dataset, I can run it normally and get the correct results, I don't know what's wrong