Random genoic amplification for microarray - (Aug/23/2005 )

My planned experiment involves chromatin immunoprecipitation followed by random amplification of the precipitated DNA and microarray hybridization.

It looks like there are two strategies for doing random amplification; the first one involves blunt-ending, adaptor-ligation, and PCR, while the second one involves initial extention from a tagged random-primer followed by PCR.

I'm inclined to go for the second method because it looks way simpler (and I've seen some very nice published papers using this method), but I wonder if there is a particular reason why some people rather prefer the first method.

Any comments or suggestions would be appreciated. Thanks.

random primed amplification can bias towards repetitive elements or certain parts of the genome that will affect how the signals are interpreted on the microarray, although I don't think this has been shown in the literature and is more anecdotal, certainly in our lab.

adaptor mediated PCR amplification does not really have this bias and actualy amplifies all fragments evenly. Our lab prefers adaptor PCR for random amplification, once optimised it is not at all more diffcult than random primed.

Nick

That's interesting...
I want to ask you a couple more questions Nick, especially because you look like a very wise man (judging from the photo).

1. Can you think of any reasons why random-priming ends up in preferential amplification of repetitive sequences? Because I don't see why, logically speaking, it should be the case.
By the way, just to make sure we are on the same page, when you say "evenly" you mean "proportionally" right? Because repetitive elements SHOULD be much more abundant than single-copy sequences in the non-manipulated genomic DNA to begin with.

2. The protocol for the adaptor-mediated method I've come across involves initial T4-polymerase treatment of DNA for blunt ending. Is this necessary if my DNA has been fragmented by sonication? Doesn't sonication mostly generate blunt ends?

I ain't wise, I just like yoda, he is so cool!

I do mean proportionally yes! As for the reasons the random nature of the hexamers may not necessarily be random at all. I have seen in some reactions specific bands within the smear, or brighter parts within the smear that suggest that there was perferential amplification of some regions within our genome (we are working with human/CHO hybrids).

Another thing that we saw with random priming is the preference for smaller fragments ie between 100bp to 2kbp and not much above this , does this affect the proportionality of each fragment within the gemone? we think so. Adaptor ligation mediated PCR has not shown this, there is a even smear that suggests a representative whole genome amplification.

I don't think Sonication gives rise to blunt ends only. It would not hurt to fill in the gaps with T4 pol in case of this.

good luck!

Nick

For those of you out there who are interested in this topic, I found this informative paper.

Liu, Schreiber, and Bernstein (2003) GMC Genomics 4:19
Development and validation of a T7 based linear amplification for genomic DNA