Removal of retained introns / primary transcripts from de novo RNAseq assembly

[sandmann]

Dear all,

I have assembled a transcriptome from > 100 Mio 2x105 bp reads and obtained a nice set of sequences without a reference genome.

Some of them are very long (> 20 kb) and a blastn/blastx search revealed that they seem to correspond to transcripts that have retained intronic sequences. [I suspect that the velvet/oases assembly is most likely correct, as unprocessed transcripts are known contaminants of polyA-selected libraries. I don't have a reference genome to proof it, though.]

Still, I was wondering if anybody had any idea on how to remove such primary transcript sequences from my output before translating the sequences into putative proteins ?

Thanks for any input,
Thomas

[tboothby]

I am also trying to identify ways to parse out intron containing sequences from RNAseq data in an efficient way. The best way I can think of is to make contigs which include both intron containing and intron lacking variants, and then remove the intron containing transcripts from your sequence database.

Unfortunately, I am yet to find a good way to do this without a reference genome.

If you find a way to do this please let us know.

Cheers,
T