Retrieve information about pseudoreplicates composition

[celinech]

Hi, nice work on the package!
Later on in my analysis, I need to get the composition of each pseudoreplicate.
I want to do a heatmap of the differentially expressed genes (DEGs) between 2 conditions, I can't do it per pseudoreplicates, but only per cells in my data, while the DEGs where identified through pseudoreplicates.
Do you think it would be possible to add an output where the user retrieve the composition in cells of the pseudoreplicates ?

An additional remark is that you could add a seed to reproduce the pseudoreplicates generation.

Thank you again,
Cheers,
Céline

[ChristophH]

Hi Céline,

Thanks for the suggestion. I can see how in some cases it might be useful to know which cells were pseudo-bulked - that's something we should be able to return somehow.

Regarding the RNG seed, I'd rather not mess with the RNG and instead leave that up to the user. Nothing stops you from using set.seed() before calling a DElegate function.

[celinech]

Thank you for your feedback, and for taking into account my suggestion.

As I already use your library in my pipeline, with the set.seed() before calling the findDE function, I tried to retrieve the pseudo-replicates by downloading the library as a .zip archive, modified it to get the pseudo-replicates counts matrices and recompressed it. I did the library installation through the .zip file and ran my pipeline (still with a set.seed()).

Interestingly, I couldn't reproduce the results I had before, based on the fact I couldn't reproduce the same pseudo-replicates as before, even with the use of the set.seed() function before calling the findDE function.
Would you guess what would be the source of the differences I get?

Many thanks,
Céline