DEG pattern while using unspliced and nuclear counts

Hello everyone,

I have some basic questions regarding the differential expression results obtained from nuclear counts and unspliced counts in single-nucleus data.

For instance, certain groups of gene expressions are upregulated when computing DEGs using nuclear counts, while the same gene expressions are downregulated when using unspliced counts. I am curious about the biological meaning behind this discrepancy.

From my understanding, nuclear counts primarily capture the transcriptional dynamics, whereas unspliced counts represent changes in pre-mRNA levels.

I would greatly appreciate your advice.

Thank you,