Low percent alignment - Hiseq

[Jin1]

Hi All,

I just got my first batch of HiSeq data, but it seems not so good. We are aiming at collecting 3' junction information from Arabidopsis cells, and there's 99% of my raw reads including polyT. However, the sequences after trimming off the polyT residues turns out to be very low in quality (see the attached image) and only 13% can be mapped to the genome.
Since we are constructing this library from very low amount of cells, we did the first RT with T7-oligo dT primer and applied T7 in vitro transcription to obtain enough amount aRNA for following cDNA library construction. We sheared the aRNA in the RT buffer at 94C for 6min and then did adaptor ligation.
I am now wondering why the low percent alignment due to the experimental / library construction steps, or just sequencing problem? I'm planning to retry the T7 and RNA shearing steps to increase the length / detect sequences of insert part (although we did small-scale sequencing and QC for this library). I'm guessing it might get a better align rate if we increase the length of insert. Anyone has good suggestions?
Thanks!

[Brian Bushnell]

I don't really have any suggestions for library prep or insight into why the quality is low, but you should get a higher mapping rate with BBMap as it's geared toward extremely low quality reads.

The defaults should be fine, but you can get a higher mapping rate with these flags:
"k=12 minratio=0.3 slow maxindel=4000 local"

-Brian

[nucacidhunter]

There is not enough info on library prep method after fragmenting aRNA to work out posssible causes.

[Jin1]

There is not enough info on library prep method after fragmenting aRNA to work out posssible causes.

Yeah, that's exactly my concern. I don't know how to optimize this. Just total blind guess...