Output RNA-seq reads that don't map to the transcriptome

Question

[arturo.didonna]

Hi everyone. I am trying to identify reads from human RNA-seq data the are NOT mapped to the transcriptome. Is there a way to get a software like SoapSplice to output this kind of reads?

Thanks in advance.

Arturo

[Manvendra Singh]

Once you have transcriptome ref-seq and its bowtie index, run tophat2, and then from output unmapped.bam is consisting the reads you are interested in , isn't it?

Answer 1

[Devon Ryan]

Why not just use STAR/tophat/whatever to map as normal and then use htseq-count and parse the output of the -o option? You'll then have all alignments annotated with whether they align to a feature (just grepping for something like "no_feature" should then find all of the alignments you're after).