Inquiry: minimum length of reads for referece-based assembly or de novo assembly

[sunfuhui]

Hi all,

While doing QC preprocessing of RNA-seq reads, what could be the minumum length of reads for a reliable assembly: reference-base or de novo assembly?

I am a greenhand in RNA-seq area, and just wonder, when we met bad quality data, to what extent could we trim the reads?

for example, as fastqc results shown in the attachment.
Do I have to trim the first 10bps and the last 20bps?
Is 50bps left enough for (tophat->cufflinks) assembly?

Thank you very much in advance!!!!!!

[persorrels]

Did you get a new insight on this problem? I posted a similar question here:

http://seqanswers.com/forums/showthr...919#post117919