could scde use in differencial expression transcripts analysis?

[DamienTan]

Dear scde team,
Thank you for this great scde package! When I read the tutorials Single-Cell Differential Expression Analysis, I noticed that it is said

The analysis starts with a matrix of read counts. Depending on the protocol, these may be raw numbers of reads mapped to each gene, or count values adjusted for potential biases (sequence dependency, splice variant coverage, etc. - the values must be integers).

I want to know if a trancript count matrix could as an input file for scde instead of a gene count matrix. Is the transcript count matrix still obey the same distribution as gene count matrix (negative binomial distribution)?
Look forward to your reply! Thanks :)