reliability for the usage of merged GTF from tophat-cufflinks as an novel annotation?

Question

[Gabby]

Hi! I want to study alternative splicing events between two different conditions using RNA-seq data of new paramecium strain [lets say "A"]. I want to use MATS for this using merged gtf annotation from tophat and cufflinks, which is based on reference and annotation(GFF3) of another closed related strain["B"] of paramecium.

My question is:

[1] Is it okay to use this merged GTF for this analysis? (I found difference on IGV viewer between merged GTF and GFF3 of "B strain" eg. different gene names for particular coordinates of reference of "B").

[2] I am also looking for another approach/s, which is better for this study?