Finding out what sequence to study by ChIP - (Jul/28/2008 )

Hi everone,


I am completely new to this field and I got one million questions to ask

But most importantly, I keep thinking about one thing (which is quite philosophical):

How people figure out what region in the DNA to study???

Say for example, I want to know whether a gene "A" has changes in histone modification in a certain type of cancer

And so I have some samples of liver cancer and normal tissue

I hypothesis that gene "A" is turned on in liver cancer

So I want to see if the histone near the promoter of gene "A" has more H3K4-tri-me and less H3K27-tri-me in cancer compared with normal tissues.

Now there's a problem....WHERE should I look???

For example, after I do CHiP using H3K4me3 and H3K27me3 Antibody to enrich the DNA, what should I do next if I don't know WHERE in the promoter (or even more upstream) to LOOK?

Design primers for PCR every 100bp? or do Southern??

Can anyone help me to understand more about the experimental flow???

---

I am not sure if I will be of any help....but..

We are doing a similiar thing, but we're going 'all the way' and doing chip-chip. That gives us regions which are enriched (using anti-histone antibodies). Then we design primers that are specific to these enriched regions.

I know someone in our department that is looking at specific genes. She got some some slides made up that only contained her gene of interest - she did the chip-chip a couple of times to find which area was interesting.

Arg, it's too hot and my brain is melting!

Clare

---

If I have no genes on my list the best thing would be Chip-Chip,

but In fact I have about 10 genes to study....

It sounds like I can make my own slide that only target the promoter of the genes I want??

How could this be obtained?

---

If I have no genes on my list the best thing would be Chip-Chip,

but In fact I have about 10 genes to study....

It sounds like I can make my own slide that only target the promoter of the genes I want??

How could this be obtained?
[/quote]

You could easily get a slide made up with only 10 or so genes on it. I am using a custom made chip (I put almost 200 genes on it, which is damn expensive!!!). I don't know where you are, but I'm sure there are some companies around. I use Agilent but our microarray facilities also print slides.
http://www.chem.agilent.com/Scripts/Generic.ASP?lPage=43127

---

Your information is really helpful

Do you know how much is a custom chip if I am only studying 10 genes?

---

Your information is really helpful

Do you know how much is a custom chip if I am only studying 10 genes?
[/quote]

Sorry, I have no idea! It would depend on the company. For example: Agilent uses a really fancy method and so our slides are about £300 each. You would just have to search around.
Good luck

Clare

---

Thanks a lot, I will consider that!!


Just thinking about alternative options, what about this:

Sonication the DNA into fragments of about 1-2kb

Then precipitate with H3K27me3 Abtibody,

for each gene, design 3 pairs of custom primers (amplicon about 100-200bp), which amplify different regions in the promoter and exon1 (eg. primer pair 1 at -1000 position, pair 2 at -200 and pair 3 at +100)

Then do realtime qPCR


Will this work??

---

You could do that - Although I would sonicate your DNA to smaller sizes. We average 200-600bp.

Just out of interest - what genes are you looking at? We are doing a H3K27me3 ChIP-chip and perhaps your genes are on our slides (I mean, the odds are probably low, but who knows!).

Clare

---

I am curious about the size after sonication....

Why 200-600 is better?

(cos I know very little about chip)

if the size is too small, there's a chance that the PCR wouldn't pick up the fragment, and I guess too big the fragments will also generate problems (although I don't know what the problem will be)

Can you explain a little bit about choice of DNA fragment size?

---

Can you explain a little bit about choice of DNA fragment size?
[/quote]

Well, as we are hybridising our ChIP DNA to an array we want to minimise non-specific binding. For example: say if our histone binds to a region which is 300bp, and yet our DNA fragments are 2000bp long, that gives an extra 1700bp of DNA that could hybridise to our array but isn't actually bound to the histone.

Does that make sense? I have been drinking wine whilst packing in preparation for moving house. Ahhh bioforum gives me a nice break

Clare

---