We detected sub-clonal poly(A) truncations in mRNAs transcribed from plasmid templates using this method, and have had more consistent results with PCR-based templates, as per @sethcheetham.bsky.social's reply. Dorado poly(A) analysis needs to specify any non-A linker sequences to be accurate.

We observed ~20% deletion error rates in the poly(A) tails of our plasmids when we sequenced using RBK114 (Rapid) kits. This is pretty approximate though - direct RNA or cDNA sequencing, then poly(A) tail length analysis with Dorado is more reliable (it corrects for homopolymer errors) but $$.

I would also like to know if you prepared your library from linearised or supercoiled plasmids? Did you use a rapid or ligation sequencing kit? And did your poly(A) tail incorporate a (non-A) linker sequence? Sorry for the 20 questions - they all count!