MeDIP Seq - Peak Strand

[saint_667]

Hi guys

I have been working with MeDIP-Seq data and one of the problems puzzling me is that:
If we have a peak as reported by the peak caller in a location on the genome, how do we know that this peak originated on the plus strand or the minus strand.

http://www.nature.com/nature/journal...re10008-s1.pdf (Figure 10) shows the mechanism of how one can conserve strand information - however where I am confused at the moment is that after our paired end sequencing run we have two files one with /1 and other /2 parts of the paired end reads.

Bowtie e.g. will take a read from /1 and a matching pair from /2 (reverse complement it) and try to align it to the reference genome. How can we figure out that this read originated from the + or - strand in?

Any help here would be greatly appreciated.

Thanks

[dpryan]

The strand of the original fragment is equivalent to the strand of read #1. Note that whether strandedness is conserved is dependent upon how the library was made.

[fkrueger]

The orientation of a read pair is always determined by read 1 of the pair. By definition, read 2 has to be the reverse complement or read 1 as it is regenerated by PCR. The MeDIP pulldown precipitates the strand that has the methylation signal, and this strand will always carry the read 1 adapter at the start (sort of illustrated in Supp. Fig. 10). hth

[saint_667]

Thank you for clearing this up

So if I wanted to find if the peak is from the + strand or the - strand, I can perhaps look in the samflags and split the sam file it into 2 bam files depending on which strand the read 1 (read /1 from the FASTQ file) of the pair originated from?

Thanks again

[dpryan]

Yes, that'd work.