Anyone using Bowtie2 for ChIP-Seq?

[blancha]

Hi,

Is anyone using Bowtie2 for ChIP-Seq?
Everyone seems to be using Bowtie 1.
I have no one with whom to discuss the best settings for ChIP-Seq with Bowtie2.

I just align the reads with the default settings for Bowtie2.
I've already completed several analyses, and the results seem correct, but I have nagging doubts that I'm not doing it properly.

For example, those using Bowtie1 often put the parameter --best.
Should I be filtering my alignments after running Bowtie2 to keep only best alignments?

Any feedback is appreciated, even if it is to tell me that I should just be using Bowtie1 like everyone else.

Thank you,

Alexis

[ewels]

Hi Alexis,

Typically, we choose whether to use bowtie 1 or 2 depending on the read length of the library. As ChIP-Seq libraries are often sequenced using relatively short reads (around 50bp or less), bowtie 1 is probably more suitable for alignment.

The --best parameter isn't really applicable in bowtie2 because it handles alignments differently to bowtie1. The defaults in bowtie2 are pretty sensible, so if your results look correct then you're probably fine.

Phil

[blancha]

Great.
Thanks for your answer tallphil.
I feel better that I'm not delivering results that are completely false.

I've been using bowtie2 simply because 2 sounds better, and bowtie2 can handle compressed FASTQ files directly.

My reads are paired end 50 bases. The documentation on the bowtie2 website is a bit ambiguous since they recommend bowtie2 for reads longer than 50 bases, and say that bowtie1 may or may not be better for reads shorter than 50 bases. I may be a bit picky but they don't say anything about reads exactly 50 bases long.

I should probably switch to bowtie1 for my 50 bases long paired end ChIP-Seq, since it seems to be much more widely used that bowtie2.

It would be interesting to do a comparison of the 2 programs on 50 bases paired end ChIP-Seq, but since the programs are constantly being updated the comparison would have to be repeated for each update.