Bowtie2 / eXpress

[romerosoriano]

Hi,

I was planning to use Bowtie2 to align RNAseq reads to a list of transposable elements and I wondered which could be the best option for treating multiple alignments (as my fasta file contains distinct copies for each same repetitive sequence). I usually use Bowtie2's "-a" option, which makes the program search for and report all valid alignments, then I use eXpress for the subsequent expression quantification (letting eXpress "decide" where my multiple aligned reads belong to). Having to deal with repetitive sequences, I think that maybe "-a" option will lead to a huge alignment file, and I was wondering if using Bowtie2's default options (report the best alignment) would be more straightforward.
Does someone have experience with these options? Is eXpress substantially more accurate than Bowtie2 when finding best alignment?

Thanks

[peromhc]

The eXpress model explicitly requires multiple mappings be included, so you have to use either -a, or something like -k 20. The latter actually, is the preferred method.