I have discovered new transcripts via TUXEDO RNAseq pipeline. After running cuffmerge I have got a GTF file with genomic coordinates (Ensembl build 37) of individual exons of discovered isoforms. I want to verify their existence by PCR approach. Designing isoform - specific primer sets means I need their sequences. Can anybody advise on how to obtain full transcript sequences? Note that I am not worried about getting these sequences directly from RNAseq reads (not sure if that's possible or not; if possible an explanation of how will be appreciated though) but rather want to reconstruct them using GTF annotation and reference genome sequence in high throughput manner. The high throughput aspect is important.