Perplexed!

[arkal]

Ok so this isn't really a problem per-say but i was just curious as to what it really means....

I ran picsrd's insert size metrics on my PE RNA Seq data (mapped to the Ensembl Transcriptome model) and iin the results i get the size of largest insert as 11kb.....

does this mean that a read in a pair has mapped to 1 point and its mate has mapped to another gene somewhere else in the consolidated transcriptome reference, 11kb away, or does it means something else altogether?

[dpryan]

What makes you think that 11kb would indicate a different gene. I'm not sure how informative the insert size metric would be on RNAseq data, given that it's going to be spliced. 11kb could easily be the next exon. Just as an example, I have single reads that span 15kb, simply because of splicing.

[arkal]

HEy thanks for your reply! Did you map your RNA Seq data to a trasncriptome reference or to the HG18/19 Human Genome reference?

I've mapped my data to the Ensembl Transcriptome reference so shouldn't the introns not be present in it at all? So there's no question of it being so far away due to splicing.

[dpryan]

Generally you would map things to the genome with a splicing aware aligner (tophat2/bowtie2, gsnap, etc.). What makes you think that Ensembl has all of the existing splices for a gene? It's quite possible that you have reads covering splicing events that are unannotated. It's also possible that you just have on read misaligned. It's often helpful to look at the data in IGV or similar to get a better feel for these sorts of things.

[arkal]

You got a point there! It's not 100% only exons and there may be unannotated regions. I hadn't considered that.

And yeah i'm running tophat now. I ran a BWA of my reads onto the trscpt ref first to see what sort of duplication rate i'm really looking at.

will check out the alignment on IGV asap!

Thanx a ton!