Trouble plotting heatmap of correlations based on sample conditions

[Jenwie]

Hi!
I want to create separate heatmap_compare_corr plots based on the sample conditions that I included in the "interaction" model. As is, with the heatmap_compare_corr function this is not very straightforward and it's not possible to easily add titles or customize anything about the complex heatmap that is being drawn. I was wondering - would it be possible to return the heatmap or return the matrix itself in this function such that there is more flexibility when plotting the different correlations? Right now, the function just draws the heatmap and doesn't return it.
Thanks a lot in advance!

[astaburuaga-garcia]

Hi Jennifer! Thanks for your message.

Thank you for your message and for raising such an insightful point! You're absolutely right about the need to separate heatmaps by sample when using the interaction model.

I'm pleased to share that we've implemented some helpful improvements based on your feedback:

1. The heatmap_compare_corr() function now returns a heatmap object that can be stored for further use:

heatmap_corr <- RUCova::heatmap_compare_corr(
  sce[,sce$sample == your_sample_name], 
  name_assay_before = "counts",
  name_assay_after = "counts_interaction", 
  name_reduced_dim = "PCA"
)

2. We've added compare_corr(), which underlies the heatmap function (so it needs the same arguments), to directly export the correlation matrix for custom visualization:

matrix_corr <- RUCova::compare_corr(
  sce[,sce$sample == your_sample_name], 
  name_assay_before = "counts",
  name_assay_after = "counts_interaction", 
  name_reduced_dim = "PCA"
)

These changes should give you more flexibility in your analysis workflow. Please don't hesitate to reach out if you have any other questions or suggestions—we're always eager to improve the package based on user needs!