This is a test version of Biostars. For the public version, visit https://www.biostars.org.
bowtie problem: cannot locate bowtie index

Hi everyone,

I actually have three minor questions about bowtie. I wrote a run.sh file as follows:

bowtie2-build rrna.fa rRNA
bowtie2 --very-sensitive  -x rRNA  -f reads.fa --un a.fa
bowtie2-build genome.fa GENO
bowtie2 --very-sensitive  -x GENO  -f a.fa  --un b.fa

However, when running bash run.sh, I was encountered a problem as follows. This is my first question.

Could not locate a Bowtie index corresponding to basename "GENO"
Error: Encountered internal Bowtie 2 exception (#1)

My second question is related to bowtie --very-sensitive option. Does this option map more reads on the genome?

My third question is about bowtie vs tophat. My understanding is: bowtie can not detect exon splicing, so abandoned some mapping reads; in contrast, Tophat overcomes this problem and enables more reads mapping on genome. Is my understanding right?

THANK YOU very much!

bowtie
• 9,376 views • link
updated by Ram Ram • written by biolab biolab

1 answer

  1. Presumably something was mistyped in the bowtie2-build command when you tried to build the GENO index. Just look and see what files are in that directory and how big they are. In practice, it's simplest to just run the bowtie2-build command yourself and ensure that it doesn't throw an error message.
  2. Not necessarily. I've had it map fewer with that option, though the difference is quite small. The --very-sensitive and --very-sensitive-local presets mostly affect reads that can align with a high edit distance. However, since the seed size changes with these, you may occasionally exhaust the number of seed extensions before finding the actual best alignment (this is true with any setting, but is more likely to occur as the seed size is decreased). Having said that, you'll typically get a few more reads mapped with these settings, though it'll take longer to do so.
  3. Correct, bowtie2 isn't intended to map mRNA to the genome and will miss splice junctions. If you have that sort of data and sufficient RAM, then use STAR instead. It's MUCH faster than tophat2 and produces equally good results.
• 0 views • link
updated by Ram Ram • written by Devon Ryan Devon Ryan

Hi Devon,

Thank you very much for your detailed explanations, all of which are helpful for my work.

• 0 views • link
updated by Ram Ram • written by biolab biolab

I'm accepting this answer so Biostars bot doesn't bump it again.

• 0 views • link
written by Ram Ram

Log in to answer this question.

More posts like this