Protocol Online logo
Top : Forum Archives: : Macroarray and Microarray

Quality of genomic DNA for array CGH - Can "Genomiphi"-amplified genomic DNA be used? (Mar/15/2006 )

Hi,

I was wondering if it´s possible to use genomic DNA in an array CGH experiment, where the DNA has been amplified by a "whole genome amplification" kit, such as Genomiphi by Amersham/GE Healthcare.

I have a very limited amount of patients DNA that I would like to compare to "normal" DNA by means of Comparative Genome Hybridization. In order to increase the DNA amount I used the Genomiphi kit, now I don´t know if it makes sense to use it. I´ve read the kit is based on hexanucleotides used as primers for the polymerase, so the distribution of annealing sites makes a difference concerning copy number, right? Or am I wrong. Please tell me your ideas.

And: Is Genomiphi-amplified DNA usable for Southern applications?

Thanks in advance.

-FMM-

QUOTE (FMM @ Mar 15 2006, 10:45 AM)
Hi,

I was wondering if it´s possible to use genomic DNA in an array CGH experiment, where the DNA has been amplified by a "whole genome amplification" kit, such as Genomiphi by Amersham/GE Healthcare.

I have a very limited amount of patients DNA that I would like to compare to "normal" DNA by means of Comparative Genome Hybridization. In order to increase the DNA amount I used the Genomiphi kit, now I don´t know if it makes sense to use it. I´ve read the kit is based on hexanucleotides used as primers for the polymerase, so the distribution of annealing sites makes a difference concerning copy number, right? Or am I wrong. Please tell me your ideas.

And: Is Genomiphi-amplified DNA usable for Southern applications?

Thanks in advance.


Hi FMM,

The short answer is yes, WGA products work very nicely for CGH. The amplification is highly linear, with very limited bias between different parts of the target sequence.

Sequence bias in any 'generic' amplification method can arise from primer-binding bias and also from preferential amplification of some sequences by the polymerase used. In WGA, the polymerase (Phi29) is highly processive, so it's unaffected by things like secondary structure (which might cause lesser polymerases to 'fall off' the template strand). The random hexamers used will anneal about every 500bp in a random DNA sequence, so the effect of primer-binding bias over a whole genome is effectively zero.

You should normalise the amount of starting DNA for the WGA reactions, although the yield of amplified DNA is highly reproducible if you run the reactions to completion (18 hours, as recommended by the manufacturers). The amount of product is limited by the dNTP concentration, so the reaction stops when they run out.

You will almost certainly need to fragment the WGA products (short DNAse digestion) prior to array hybridisation. For a 20ul WGA reaction (after heat inactivation at 65oC for 10min) I add 1ul of 0.1U/ul DNAseI (NEB) and incuabate at 25oC for 2-3 min then immediately heat inactivate again (95oC, 10min). This gives fragments in the 100bp-1kb range with my (prokaryotic) genomic DNA.

In principle your fragmented WGA products should also be fine for Southern blotting, although I haven't tried this.

hope that helps,

D.

-del-

Hi del,

I did not expect anyone to answer that quickly, so apologies for the delay. Thank you very much for your answer, your help is very useful, since I´m running out of DNA and still might want to perform CGH.

One more question arose when I read your answer: You said the primers bind about every 500bp, but what about repeat regions. They´ll probably be amplified as long as any primer binds right next to it, but the larger the repeat region the less the probability it´s being properly amplified, isn´t it?
Wouldn´t that be a problem in CGH? And in Southern hybridization, for example in search for translocation breakpoints, where breakpoints often will be found in regions with many repeats.

I´m very much hoping to read your ideas. And again thank you very much indeed for the help so far.

FMM

-FMM-

Hi FMM,

QUOTE (FMM @ Mar 21 2006, 08:04 AM)
One more question arose when I read your answer: You said the primers bind about every 500bp, but what about repeat regions. They´ll probably be amplified as long as any primer binds right next to it, but the larger the repeat region the less the probability it´s being properly amplified, isn´t it?
Wouldn´t that be a problem in CGH? And in Southern hybridization, for example in search for translocation breakpoints, where breakpoints often will be found in regions with many repeats.

I suspect this wouldn't be a problem, for several reasons. Because you're priming with a population of random hexamers, there will certainly be primers in the mixture which will anneal within the repeat regions. More usefully, Phi29 polymerase apparently extends up to 60kb from a single priming event (with multiple subsequent priming events on the newly synthesised strand). This means that you should be able to reliably amplify through repeat regions from priming events which might be some distance up or downstream. You might see a small amplification bias, but this hopefully wouldn't be too serious. From the published work using Phi29 the bias between different genomic loci seems to be only 3-4 fold at most, and normally much less. If anything, you might expect the presence of really long repeats to titrate out some primers from the random mixture, which could conceivably (maybe...) affect the amplification of other sequences elsewhere in the template. However, the primers should be present in large excess, which will avoid this, especially if you limit the amount of input DNA (don't add too much - less is more!).

You could always do it as an experiment, and try and estimate the amplification bias (if any) between a couple of different loci which are inside vs. outside the repeat regions in your genomic DNA, if you're worried about the impact on your CGH or Southern results. Some qPCR work (or PCR on end-point dilutions) with your DNA before & after amplification should do it, but I guess you probably have enough to do...

D.

-del-