Bowtie alignment using RY format as reference

Question

[Nalini]

Hello! I have E.coli ABI-SOLiD data downloaded from NCBI SRA in the fastq format. I want to now align this on Bowtie-2 using a reference genome where R can be any purine and Y could be any of the pyrimidines. So I would align the fastq files to a reference file1 containing ATGY ( where I want Y to align to either C or T ) and another reference file2 which would contain ATCR ( where I want R to align to either A or G). My question is whether alignment on Bowtie with such a reference sequence is possible on Bowtie or do we have to specially instruct the system that R=A/G and Y=C/T? If the latter, how is that done? If Bowtie doesn`t permit this, is there any other tool that can be used for the same purpose? Thanks!!

Answer 1

[Devon Ryan]

1. Bowtie2 doesn't do that. I'm actually not aware of one that supports all IUPAC ambiguity codes like that.
2. It sounds like you're aligning BS-seq data of some sort (though I didn't think E.Coli had methylated DNA), in which case you should use an aligner for that type of data.
3. You're better off staying in color space than going to a non-colorspace fastq.

[Nalini]

Thanks for your reply. Clarifications point-wise:

1. Okay. Could I try Novoalign. I was reading that they accept IUPAC formats. Any suggestions on that?
   1. Yes, these are BS-treated samples. Purpose is to look out for R-looped regions in E. coli by looking for C to T changes which would occur on the single strand of the R-loop.
   2. Yes, I was intending to use csfasta, but somehow couldn`t find that option in NCBI SRA while downloading the reads.

[Devon Ryan]

1. Sure, I'd forgotten about that one. It'll be slow unless you're paying for a license, but presumably it'll do what you need.

Anyway, you'll be best served by either bismark with bowtie1 or maybe novoalign, which I think has a bisulfite aligner that also support colorspace data. For the SRA data, you're looking for abi-dump.