Forward strand bias?

[dawe]

Hi all,
this may be something known... anyway, tags coming from Illumina sequencing method (GA II here) seem to map preferentially on forward strand... I haven't check if there is a statistically significant variation, but apparently tags aligned on forward strand are 2-5% more (in various ChIP-seq and also in genome resequencing).
I wonder which could be the source of this variation... The library prep? The cluster generation? The aligner? None of these steps makes sense to me...

d

[steven]

?! strange indeed!
statistical significance may be an important point here.. is the bias constant across chromosomes?
s.

[bioinfosm]

What aligner are you using? Probably you could use a different one and see what happens to the bias

[maubp]

I'd also suggest taking your references and reverse complementing them. Then repeat the alignment on the RC reference and see what happens. If this is just a software artefact, then you should see more reads on the new forward strand (i.e. what was the reverse strand in the original reference).

[steven]

My bet would be a single genomic position (on the forward strand) that gets a gazillion of copies of a single read (generated by a contaminant, an amplification artifact, an incorrect filter, or something else).

[steven]

Hi all,
this may be something known... anyway, tags coming from Illumina sequencing method (GA II here) seem to map preferentially on forward strand... I haven't check if there is a statistically significant variation, but apparently tags aligned on forward strand are 2-5% more (in various ChIP-seq and also in genome resequencing).
I wonder which could be the source of this variation... The library prep? The cluster generation? The aligner? None of these steps makes sense to me...

d

Hi,
any news regarding this issue?
thanks,
s.

[dawe]

Hi,
any news regarding this issue?
thanks,
s.

Not yet... I've been busy on other issue but I will investigate this further...