using RNAseq reads to verify predicted gene models

Question

[sullis02]

this was touched upon in https://www.biostars.org/p/275944/ but not relevantly for me.

BRAKER, when run on the same genome more than once, with the same parameters, produced many duplicate models, but also some different models for the same gene start/stop coordinates. It also produced some gene predictions that are nested (contained within) a gene predicted in a different run.

I would like to use RNAseq data to decide in such cases which of these models is more likely to be correct. (Realizing that this will only help if the gene is expressed)

Are there tools for using RNAseq data to define the boundaries of exons/CDS within a predicted gene feature?

I am familiar with STAR for mapping RNAseq data.