negative control or not

[ronindan]

hi, everyone, I am a new one in this area.
now I plan to use this method to study a TF binding sites.
I prepare two biological replicates, but whether I need a negative control sequencing?
I read some paper, and say there are two case can choose:
one-sample analysis where only a ChIP'd sample is sequenced
two-sample analysis, where both a ChIP'd sample and a negative control sample
I was a little confused and want to hear your opinion.
thanks very much

[kopi-o]

You should definitely have a negative control for ChIP-seq.

[ETHANol]

Do the negative control (normal rabbit serum), if your ChIP goes well the negative control will not have enough DNA to sequence as determined with QuanIT HS (QubeIT) DNA assay kit.

You absolutely have to sequence your input DNA alongside your ChIP'ed DNA. That sets your baseline.

[ronindan]

thanks to both of you
I have more questions.
You known, I am working on a rice transcription factor, in order to do the ChIP-seq, I fusion the Flag tag with my TF and transformed into the mutant background. So I have two choice about the negative control:
1.mutant sample with Flag anti
2.transgenic sample with no anti
which one is better?
by the way, how many biological replicates should I seq in my anti sample? two is enough?

[mudshark]

provided you get enough/comparable amounts of DNA pulled down i would prefer option 1 (mutant sample with anti-flag).

just to get it right: your intention is 'just' to map transcription factor binding sites?
if yes, i would simply sequence the input (and not the mutant sample with flag anti). reason: you can use the input to control any other ip with different antibody. the mutant sample with anti flag is only good for comparison with your flagIP.

what do you define as biological replicate? also different transgenic lines? will the expression of your construct vary between replicates? if yes, replication is a good idea. if no, i would not replicate in the first place. one sample, have a look at the result, check some loci with qpcr compare with literature.. and maybe do a replicate later. often a single expt is enough for most of the mapping.
maybe you also should consider pooling of replicates before sequencing if there are strong variations expected.

[ronindan]

got it, many thanks of mudshark
all of your answer were helpful

[ETHANol]

I assume you are using a FLAG-tagged approach because you do not have an antibody to your protein of interest. If that is the case then I strongly recommend you produce your own antibody. It is a easy and simple thing to do and the cost in terms of labor and money is minimal in terms of the cost of the ChIP-seq project. Yes it takes 90 days, but too often I see people two years into a project and still trying to get by with no or a low quality antibody. How to make good antibodies is a whole different topic so I'll leave it at that.

As far as your last question, I don't understand what you mean by 'mutant'.