Selecting relevant binding sites according to MACS data

Question

[robbenstijn]

I've got ChIP-seq data analyzed with MACS and looks at such:

> chr   start   end   length  summit  tags  -10log10(pvalue) fold_enrichment     FDR(%)
> chr1  9911    10616  706     251     77       60.46               4.07         0.30
> chr1  11119   11740  622     185     32       101.80              10.07        0.13
etc...

How do I now select the relevant peaks according to these scores. I'm looking for a sort of threshold-value for which I can say "Any peak with a score above this value is an actual binding site".

I read the Zhang article already but couldn't find any parameters to use.

Many thanks!

Stijn

Answer 1

[le2336]

If you have biological replicates, you might want to use MACS2 in combination with the Irreproducible Discovery Rate (IDR) pipeline to identify peaks that are reproducibly detected between your replicates. If you only have a single replicate there is a workaround discussed here, though replicates are highly recommended: https://groups.google.com/forum/#!searchin/idr-discuss/no$20replicates%7Csort:relevance/idr-discuss/edIg7xjism8/B4XsmTQJZRsJ