De novo alternative spicing

[Zhenhua Wu]

For PSG, it requires a known PSG model constructed from a known transcripts annotation file such as a GTF file. I am interested also in new transcripts (new exon, or new isoform), in that case, I would think to use an assembler (such as cufflink, trinity etc.) to generate a gtf file based on the rnaseq data before runing psg. Do you have any recommendations for de novo isoform discoveries using psg?

Jeremy