I'm sure that the anwser is obvious for some of you but I just endend my internship where we found novel splice junctions (junctionseq; RNAseq data) (see the attached file) and I have some difficulties with splice junctions. I understand that reads can map in the junction of 2 exons (which can be linked to quality assesment) and that comparing 2 conditions, we can find groups of exons more "associated" but I dont understand :
A splice junction is formed from a pre-mRNA (or primary transcript) whenever an intron is removed, or spliced out. Two segments of RNA that used to be separated in the pre-mRNA are now ligated to each other, forming a junction of two exons.
RNA-Seq reads that span such a splice junction site would not align very well to the genomic sequence, however if you create a file of all the potential simulated splice junctions there may be some RNA-Seq reads that now align perfectly.
The distinction between what constitutes an intron, and what constitutes an exon is purely empirical. While it is true that the spliceosomal complex preferentially recognizes and binds to properly juxtaposed splice donor and splice acceptor sites, a variety of different processed, mature, transcripts can result from a single primary transcript. This differential parsing of the cis-acting sequences at the boundaries of potential introns and exons is called alternative splicing.