Analysis of ChIP RT-PCR data (View forum version)



postdoc1978

Posted 10 March 2009 - 12:57 PM

I have a question about how to analyze the ChIP data that I have. I have looked through several posts and lots of people seem to have the same question I do, but I have yet to see an actual answer. I am very new to this whole technique, so my questions will be very basic.

I want to eventually do ChIP-Chip, but first need to verify that the IP actually worked. I have run RT-PCR using SYBR green. For the input sample I diluted it so that for the IP and the input I put in not only the same volume, but the same amount of DNA. The negative control is obviously the problem, as most people have seemed to say. This sample is obviously more dilute than the others, so I simply used the same volume for this sample as well, but this would obviously have much less DNA. I am not sure if this is how I should set this up, but that's the best I could come up with. I have CN and Ct numbers now, but have no idea what to do from this point. I am supposed to figure out if the IP 'enriched' for the gene of interest, but am not sure how to go about that. Oh, and just for further information, as of right now, I am using an H3 antibody (pan), which should bind to all DNA sequences. I moving towards histone modifications, but need to know how to analyze this data first. I would appreciate anyone who can help me with this, as I have exhausted all avenues that I can find.

Thanks for any help you can give me.

jhb80

Posted 11 March 2009 - 01:42 AM

Hi -

There's a good job of explaining done by jiro_killua on the matter of calculating and normalizing ChIP data at http://www.protocol-...?showtopic=6191

postdoc1978

Posted 11 March 2009 - 08:43 AM

Thanks. I have read that post, but I am at the very beginning stages and really don't know that much about real-time, so I need something even more detailed starting from the beginning than what was posted.

Hi -

There's a good job of explaining done by jiro_killua on the matter of calculating and normalizing ChIP data at http://www.protocol-...?showtopic=6191


KPDE

Posted 18 March 2009 - 12:41 PM

I have a question about how to analyze the ChIP data that I have. I have looked through several posts and lots of people seem to have the same question I do, but I have yet to see an actual answer. I am very new to this whole technique, so my questions will be very basic.

I want to eventually do ChIP-Chip, but first need to verify that the IP actually worked. I have run RT-PCR using SYBR green. For the input sample I diluted it so that for the IP and the input I put in not only the same volume, but the same amount of DNA. The negative control is obviously the problem, as most people have seemed to say. This sample is obviously more dilute than the others, so I simply used the same volume for this sample as well, but this would obviously have much less DNA. I am not sure if this is how I should set this up, but that's the best I could come up with. I have CN and Ct numbers now, but have no idea what to do from this point. I am supposed to figure out if the IP 'enriched' for the gene of interest, but am not sure how to go about that. Oh, and just for further information, as of right now, I am using an H3 antibody (pan), which should bind to all DNA sequences. I moving towards histone modifications, but need to know how to analyze this data first. I would appreciate anyone who can help me with this, as I have exhausted all avenues that I can find.

Thanks for any help you can give me.


Don't bother trying to use the same concentration of DNA for your input, IP, and mock. Use the same volume of all three (although you can dilute the input 10 fold or so if the Ct are too low to be in the linear range of your primers) in your PCRs. The calculation that you do for enrichment however you do it (either % of input or IP/mock ratio), is meaningless in answering whether your factor binds or not.

The key is to compare your region of interest to a region where you don't expect to find your factor to bind (or at least where you expect to see lower levels). Ideally, if you can find a region where your factor isn't supposed to bind (and which you could ideally confirm by another method, but this is often not possible) then any "enrichment" that you see there should be background. True enrichment at your region of interest, then is the fold increase over this negative control region. With pan H3 this can obviously be a problem since it's everywhere except telomeres. In this case you should compare a region right around the transcription start site of a gene (preferably an active one) and somewhere within an intergenic region that you know doesn't contain a promoter or enhancer. The signal should be lower at the TSS (since promoter regions near the TSS are often nucleosome free) and higher in the intergenic region. If this is the case then you can trust your signal.

An easier antibody to look at, when diagnosing your ChIP method to see if it's working, is the antibody to H3K4m3. If you compare the TSS and 3' end of a gene, it's typically night and day. We've used antibodies from Abcam and Diagenode for this and they both get great pull down.

march

Posted 23 March 2009 - 11:03 PM

I have a question about how to analyze the ChIP data that I have. I have looked through several posts and lots of people seem to have the same question I do, but I have yet to see an actual answer. I am very new to this whole technique, so my questions will be very basic.

I want to eventually do ChIP-Chip, but first need to verify that the IP actually worked. I have run RT-PCR using SYBR green. For the input sample I diluted it so that for the IP and the input I put in not only the same volume, but the same amount of DNA. The negative control is obviously the problem, as most people have seemed to say. This sample is obviously more dilute than the others, so I simply used the same volume for this sample as well, but this would obviously have much less DNA. I am not sure if this is how I should set this up, but that's the best I could come up with. I have CN and Ct numbers now, but have no idea what to do from this point. I am supposed to figure out if the IP 'enriched' for the gene of interest, but am not sure how to go about that. Oh, and just for further information, as of right now, I am using an H3 antibody (pan), which should bind to all DNA sequences. I moving towards histone modifications, but need to know how to analyze this data first. I would appreciate anyone who can help me with this, as I have exhausted all avenues that I can find.

Thanks for any help you can give me.


Don't bother trying to use the same concentration of DNA for your input, IP, and mock. Use the same volume of all three (although you can dilute the input 10 fold or so if the Ct are too low to be in the linear range of your primers) in your PCRs. The calculation that you do for enrichment however you do it (either % of input or IP/mock ratio), is meaningless in answering whether your factor binds or not.

The key is to compare your region of interest to a region where you don't expect to find your factor to bind (or at least where you expect to see lower levels). Ideally, if you can find a region where your factor isn't supposed to bind (and which you could ideally confirm by another method, but this is often not possible) then any "enrichment" that you see there should be background. True enrichment at your region of interest, then is the fold increase over this negative control region. With pan H3 this can obviously be a problem since it's everywhere except telomeres. In this case you should compare a region right around the transcription start site of a gene (preferably an active one) and somewhere within an intergenic region that you know doesn't contain a promoter or enhancer. The signal should be lower at the TSS (since promoter regions near the TSS are often nucleosome free) and higher in the intergenic region. If this is the case then you can trust your signal.

An easier antibody to look at, when diagnosing your ChIP method to see if it's working, is the antibody to H3K4m3. If you compare the TSS and 3' end of a gene, it's typically night and day. We've used antibodies from Abcam and Diagenode for this and they both get great pull down.


Hi Postdoc1978,
have you solved your problem.I also have IP DNA and Input DNA to look for the methylated gene enrichment in the IP over the Input by PCR but do not know what housekeeping gene and negative control to use.please share if you know anything.

JPchip

Posted 29 April 2009 - 06:14 PM

If you just want to know roughly if your samples are enriched over background then run your samples input, noAb, ChIP and any MOCK reaction containing no chromatin you have without diluting them.
Does the MOCK return an undetermined Ct value or one that is lower than the IPs containing no antibody (negative controls)? Good, there is no/very little contamination.
Check the amplification plots on the machine software. Highlight all the samples you know to contain DNA. Are all the amplification curves for these samples parallel at their linear region. Good, there were no misamplifications in the assay.
Check the dissociation curves if you're using SYBRgreen, do they all indicate only one species is amplified. Good, your primers rock.
Do all the inputs return a reasonable Ct value (probably between 15 and 25, -30 if your input is low)? Good, your PCR works.
Match the negative control Ct value to the ChIP of the same sample. Are the Ct values for the ChIP samples lower than the Ct values for the negative controls?
-If Yes. Excellent. You have an enrichment, but it should be more than 1 cycle different to be convincing.
A lower Ct value means the threshold was reached earlier in the cycling program and there was more starting DNA in the sample. Conversely, a higher Ct value means the reaction reached threshold later in the cycling program and there was less starting DNA in the sample.

If you want to get a better handle on just exactly how enriched your samples are over background (I can't remember what the recommendations were for ChIP-chip, sorry) then you will need to run your ChIP samples against a standard curve (I generated one by serial dilution of a known amount of genomic DNA from the same species sonicated to a the same size range as your samples, you don't need much, expect your curve to remain in the nanogram range). There is no need to dilute the samples (input, noAb, ChIP) initially but you should check that the Ct values you get all fall within the range of the standards. Occasionally I've had to dilute inputs to get them on the curve but not very often.
The Ct values of the standards when plotted against the LOG10 of the amount of DNA in them will be a straight line with a negative gradient. This allows you to directly relate the Ct values from your samples to an amount of DNA.
Rearrange the equation Excel gives you for this straight line to solve for the LOG10(Amount DNA) on the x axis rather than the Ct value on the y axis. Thus calculate LOG10(Amount DNA) for all your samples using their Ct values.
Now reverse the LOG10 function to give the amount of DNA (rather tricky maths).
Divide the ChIP value obtained by the negative control value obtained to calculate 'fold enrichment' of your ChIP sample over background.

truecolor

Posted 07 June 2009 - 07:48 PM

simulation credit auto
:rolleyes: Thanks for a great release!

method_man

Posted 23 February 2010 - 04:33 PM

how exactly does one prepare a series of genomic standards? we just ran out of our previous stock. thanks!

(i realize that you all were discussing almost a year ago now. hope someone still replies)