[polivares]

Do you think that there may be some non-random distribution of how RNA breaks in any sample determined by secondary structure? In other words, can you find any correlation between you bias that may be related on how RNA folds with itself?

[behoward]

@polivares: I have no idea. Seems like a possibility, but I'm not sure how to infer folding patterns, especially since we don't know how long the actual fragments are... only the first 25-30 bases or so... FYI: the fragmentation method was 'metal hydrolysis.'

[polivares]

I would run different trials with tools like this with different window sizes helping me a bit with the known splicing sites :S

[behoward]

So, it seems pretty clear to me now that the towers are caused by PCR. The Illumina protocol (http://grcf.jhmi.edu/hts/protocols/m...04898_RevA.pdf) fragments the mRNA, and then amplifies it by PCR. So if a breakage happens to occur at a particular location, the PCR amplified dataset would have several copies of the same fragment. Do I understand this correctly?

Also, if this is the problem, then this presents a bit of a problem for quantifying AS events. Is there a way to PCR the mRNA before fragmentation instead?