dplyr::mutate() Error about spotlight HOT 2 CLOSED

Hi Marc,

I'm uploading the print out of the objects as a file because it is so large. I used the 10X Genomics data from the example vignette. My object "h", printed out and can be found underneath "Object:h". However, I also asked it to print out train_cell_clust under "Object:train_cell_clust" so the error must be within this.
![image]
(https://github.com/MarcElosua/SPOTlight/assets/99920448/fbd05138-9e74-4931-8cad-8a15c8f2a5e3)

ObjectPrint-SPOTlight0.1.7.out.pdf

from spotlight.

Hi @ABauer26

Thank you so much for your interest in SPOTlight! Adding the comments from your email here so if other people encounter this issue they have access to the troubleshooting. I have to say that version 0.1.7 has been deprecated now for over 1 year so I would strongly recommend installing the version in Bioconductor.

This occurs after the gene set has been prepared, after the count matrix has been normalized, after the initial matrices are seeded, and the NMF model is trained. The deconvolution of spots reaches 100%, and then the dplyr::mutate() error occurs. 

I've narrowed the error down to the specific regions of code that fail, and it seems to occur right about here:


  h <- NMF::coef(nmf_mod[[1]])
  rownames(h) <- paste("Topic", 1:nrow(h), sep = "_")
  topic_profile_plts <- SPOTlight::dot_plot_profiles_fun(
    h = h,
    train_cell_clust = nmf_mod[[2]])

I initially thought it was an incompatibility with the "dplyr" package version that I was running (1.1.3), but any versions I have tried that are above version 1.0.10 (version required and above) has not remedied the issue. 

My guess at this point is that the error is happening within the dot_plot_profiles_fun. Could you please paste here what the h and train_cell_clust you input to the function look like?

from spotlight.