New to NGS - Looking for Chip-seq project advice

[Tom_C]

My lab is new to NGS, and we are looking to map the regulon of the TyrR transcription factor in our bacterium, E. cloacae. In general, we wish to identify TyrR-DNA interactions in response to the presence or absence of aromatic amino acids, which function as cofactors for TyrR binding. This would give use 12 samples in total, a no AAA control, +TRP, +PHE and +TYR all in triplicate.

We are thinking on using Illumina sequencing and multiplexing our samples so that they can all be run in a single lane. Is there a good comprehensive resource that details the entire procedure from start to end? I have a good hold on most of the steps involved, but I do not know enough yet to properly decide on how many reads, type of sequencing, etc. Thanks in advance.

[arolfe]

I don't know of a comprehensive resource either.

Question 1: type of sequencing/read length: in general, ChIP-Seq works well with relatively short reads- 35 or 50bp should be fine. You might want longer reads or paired reads if you're interested in binding to repetitive regions that will be hard to tell apart without the extra sequence.

Question 2: how many reads: This is very hard to answer, as it depends on the number of binding sites, the efficiency of your antibody, and your desired sensitivity (binding isn't binary, so deeper sequencing reveals weaker binding sites).

We used to do mammalian ChIP-Seq experiments in a GAII lane that would give 5-10m reads. So you could multiplex your 12 samples into a HiSeq lane and get 5m reads per sample, which ought to be plenty unless you're at an extreme of one of the factors I mentioned. It's probably better to overshoot on # of reads for the first try. Then you can sub-sample your data to see how many binding events you lose and use that to calibrate the number of reads for your other two replicates.

Good luck!