FAIRE-seq optimal coverage - large genome

Question

[boczniak767]

HI,

what is the optimal coverage or number of reads for sequencing DNA isolated by FAIRE (200 to 800bp fragments).

I'm working on maize genome which is rather big 2.4 gigabases.

Best Wishes,
Maciej

Answer 1

[dariober]

There are some guidelines from encode that mention something like ~30M reads for ChIP-Seq and FAIRE-Seq (see also Landt 2012). Although these refer to mouse and human genomes, they might apply to maize as well given the genome sizes are comparable.

Anyway, these are really only approximate guidelines. In practice you could start sequencing and check the library complexity with e.g. picard MarkDuplicates to decide whether to sequence more. In any case, I think sequencing more libraries (more replicates) at moderate depth is better than sequencing one or few at high depth.

In my experience/opinion FAIRE-Seq is quite a noisy technique, depending on the peak caller and parameters you choose you can get widely different numbers of peaks (orders of magnitude different). (See also this paper A Comparison of Peak Callers Used for DNase-Seq Data). For this reason, I would say that asking the question "where are the FAIRE sites" is not very meaningful. Better to ask "which sites are different between two or more conditions" in this way you cancel out the effect of the peak caller and the artifacts of the technique, which you could assume they are the same across conditions.