I am dealing with RNA seq data and i am using different softwares for detection of splicing events.
I came across hmmsplicer tool. I have a question in it. It generates 2 main files namely junctions.final.bed and junctions.noncanonical.bed.
The 5th column is the score column in the bed file. This column actually tells us the depth of that junction meaning how many reads actually span the junction. The score here includes decimal points. Now my question is that is it ok to have score/depth in decimals because tophat gives integer values and not decimals.
Moreover in 4th column sometimes i find SRR036966.9097539|junc=3 . Now wat is the significance or meaning of junc=3.
The score column is actually more complex than just the number of reads supporting the junction. It's a score that can be used to rank how 'believable' the junction is. The score takes into account not just the number of reads supporting the junction but how those reads fall across the junction (an evenly divided read is more believable than a read that is mostly in one exon with only a few bases on the other side of the junction) and the quality scores of the reads.
The "junc=3", "junc=18", "junc=2" at the end of the read is actually the number of reads supporting the junction. If you are interested in just that data, you can parse it out of the title.
