DNA amplification for chip-on-chip - WGA or LM-PCR or others? (Oct/31/2007 )
Hi everyone,
I'm new to chip-on-chip and I've sorted out my sonicating and ChIP quite nicely but now I have to think about which amplification method to use!! (I will be working with human primary cells). I will be using Agilent arrays and they suggest LM-PCR (which seems like such a long-winded method and quite a hassle). But I have been doing alot of reading and some people suggest WGA is better.....and then other papers use T7! ARG!
Does anyone here have any experience with trying different amplification protocols for chip-on-chip? Sorry if this question has been asked already, however I couldn't find it on the forum.
Thanks in advance 
Clare
Hi Clare,
we are in the process of trying all three, in the same situation as you.
we are trying the amplifications by qPCR before using arrays.
from the qPCR:
WGA worked best and was easy to set up.
LM-PCR was fine but tedious
T7 extension was the best in theory but very tedious and failed in our hands.
We will be using the WGA on chip's very soon.
Good luck
Nick
Thanks Nick 
My gut feeling was to go with WGA but as I'm new to this I though I should ask an expert
Which WGA kit did you use? Sigma?
Thanks so much for answering, I greatly appreciate it!
Clare
we are in the process of trying all three, in the same situation as you.
we are trying the amplifications by qPCR before using arrays.
from the qPCR:
WGA worked best and was easy to set up.
LM-PCR was fine but tedious
T7 extension was the best in theory but very tedious and failed in our hands.
We will be using the WGA on chip's very soon.
Good luck
Nick
My gut feeling was to go with WGA but as I'm new to this I though I should ask an expert
Which WGA kit did you use? Sigma?
Thanks so much for answering, I greatly appreciate it!
Clare
we are in the process of trying all three, in the same situation as you.
we are trying the amplifications by qPCR before using arrays.
from the qPCR:
WGA worked best and was easy to set up.
LM-PCR was fine but tedious
T7 extension was the best in theory but very tedious and failed in our hands.
We will be using the WGA on chip's very soon.
Good luck
Nick
Hi Clare, Hi Nick,
I also tested WGA Kit, and also the Affymetrix amplification protocol for chip-chip.
After some optimization (especially cycle number and Mg-Concentration) of the Affy-Protocol works nicely, and SIGMA WGA gave good results from the beginning.
For both the maximum variation I see in real time PCR after amplifying 0,5 ng/µl sonicated total input is less then 10-fold, using 8 different sets of primers.
(10fold between the two primer sets giving the most distant ct values - compared to the median the maximum variation is less the 4fold)
However, for amplified ChIP´ed DNA the variation is a little bit stronger - up to 16 fold change between not enriched sequences. No idea why this happens.
Nick, what is your experience with variation after amplification? Do you think my results are OK, or should I try to further improve?
Best wishes,
Fridtjof
Awesome, So I have decided to purify my chip DNA with the Sigma NA1020 kit (or Roche micro kit, depending on the best deal I can get) and then use the Sigma WGA kit. I will keep you posted
HAPPY FRIDAY!!
Clare
+1 for sigma WGA.
as for the variation frijof, do you mean between replicate input samples? because we see similar variation between these, but we always compare between matched input and IP samples for our controls.
Nick
Hi Nick,
Things are allways difficult to explain in ChIP-chip, with all the different kind of controls you can have.
My problem is amplification bias between different loci in a single sample.
If amplify input DNA (i.e. genomic DNA, no IP, no enrichment), and if I then compare the represenation of different loci in the amplified DNA, I would expected the same representation for all loci (as no sequences are enriched).
However, there is allways some bias, and also some noise from the quantification by realtime PCR.
I look at 8 different loci, and for each locus I compare the ct value of input DNA vs the ct value of amplified input DNA.
After amplification the ratio between the loci with the highest and the lowest representation is 8-fold, sometimes up to 16-fold.
I wonder if this is two much noise for hybridising arrays? My enrichment of positiv controls after IP is 30-50fold, so it is higher then the noise.
I hope I was able to explain my problem,
Thanks,
Fridjtof
as for the variation frijof, do you mean between replicate input samples? because we see similar variation between these, but we always compare between matched input and IP samples for our controls.
Nick
totally understand your problem Frijotf,
we would verify, by expression analysis with our histone Chip on Chip's.
there will be invidible variation which is smoothed out over three independent runs.
N
Things are allways difficult to explain in ChIP-chip, with all the different kind of controls you can have.
My problem is amplification bias between different loci in a single sample.
If amplify input DNA (i.e. genomic DNA, no IP, no enrichment), and if I then compare the represenation of different loci in the amplified DNA, I would expected the same representation for all loci (as no sequences are enriched).
However, there is allways some bias, and also some noise from the quantification by realtime PCR.
I look at 8 different loci, and for each locus I compare the ct value of input DNA vs the ct value of amplified input DNA.
After amplification the ratio between the loci with the highest and the lowest representation is 8-fold, sometimes up to 16-fold.
I wonder if this is two much noise for hybridising arrays? My enrichment of positiv controls after IP is 30-50fold, so it is higher then the noise.
I hope I was able to explain my problem,
Thanks,
Fridjtof
as for the variation frijof, do you mean between replicate input samples? because we see similar variation between these, but we always compare between matched input and IP samples for our controls.
Nick
sorry for dredging up up an old topic, but as to the variation between different loci in the one sample after amplification, do these loci have different GC%? Or are some in CpG islands and some not? There is a new thread about whether WGA is biased against CpG islands
http://www.protocol-online.org/forums/inde...3589&hl=WGA
and i wonder about your experiences now if you have hybridised to arrays?
thanks!
[/quote]
sorry for dredging up up an old topic, but as to the variation between different loci in the one sample after amplification, do these loci have different GC%? Or are some in CpG islands and some not? There is a new thread about whether WGA is biased against CpG islands
http://www.protocol-online.org/forums/inde...3589&hl=WGA
and i wonder about your experiences now if you have hybridised to arrays?
thanks!
[/quote]
I have completed my first hyb. It wasn't the nicest looking (as I stuffed up my hyb mix perhaps) but I did see some expected results. In my ChIP I used a H3K9 antibody and genes that I expected to be active, were - so that's a good start