Bowtie alignment for ChIP-seq: good alignment for input but not for ChIP sample

[nico_z]

Dear all,
I carried out a ChIP-seq experiment for which I have fastq files for both the input and the ChIPped DNA. As expected, I reckon, the input sample produced many more reads than the ChIP sample. The former also have a better quality than the latter. After trimming (to remove the low-quality 3' ends of the reads) I'm left with 35 n-long sequences. If I align these reads to my reference genome using bowtie (bowtie -n 3 -l 35 --tryhard --best) the % of reads with at least one reported alignment is 96% for my Input DNA but only 45% for my ChIPped DNA. Is this normal? Or am I missing something?

Thanks for your help.

Nico

[FWOS]

Can be Normal

Hi Nico,

Your results seem normal to me. On average, One can expect to map 40 - 60% of the IP'd reads. I have seen less than 10% mapping back, which was actually due to salmon sperm blockers contaminating the IP'd sample, but 45% is not too terrible.

It really depends on the mark you are looking at, and the specificity of the antibody that you used to target it. Since the Input looked good it does not seem to be an issue with fragmentation. Although you need to remember that the histone could have been damaged and/or the IP'd sample may not have been fully de-crosslinked.

Overall it does not sound like your experiment failed.

Best,
F

[nico_z]

Thanks

Thanks, that helped. After some additional analysis I found that many of the unmatched reads, for whatever reason (I'm guessing some form of contamination) nicely align to the Penicillium chrysogenum genome.