Protocol Online logo
Top : Forum Archives: : Real-Time PCR

Primer Dimer problem- Different perspective - COuld there be a treatment effect (May/30/2006 )

I have a different kind of primer dimer problem, just wanted to get a consensus from better knowing folks!

1. I have a Gene of interest X-designed primers with primer express, amplcon size -80 bp, analysed using SYBR -everything from ABI.
2. I have 2 groups- control and treatment and 1 tissue.
3. Ran a test run to see if the primers worked

1. All genes amplified, NTC did not amplify, Ct values of HK genes look similar, and genes previously tested amplified as before.
2. Melting curve shows single peak for all HKgenes and previous genes.

1. The control group alone shows primer dimer (peak around 70-75 tm) for X and another peak around 75-80 (Product?). The treated group shows a single peak at 75-80 (However there a slight bump -v.slight bump bet 70-75).

WIGO? blink.gif

Sorry am not able to attach a figure- havent figured out how to export figures from the new SDS software yet!


I have found that primer-dimers in realtime can be quite sensitive to amount of template you add; there is a really fine line between enough and too much

I would titrate your primers down and see what happens (I'm guessing you have quite a significant increase in template with your treated samples, favoring amplification and not the dimers)


I was thinking about increasing the template conc.BUt lowering primer conc is a better idea as it will save my sample some.


SO I diluted my Pirmer conc 1/2 and 1/4. At 1/4th the conc. I have a single beautifult peak for all my samples (treated and control).
However, at the original conc. Only 1 of my control samples shows primer dimers the other (I originally had 2) is similar to treated samples ( I peak with a slight bump prceding it!).
I am not sure whether there was a problem with my pipetting (I pipette 1 sample at a time)

What is your take on this?


master mixes are your friends and replicates are very important when you are optimizing, and later for statistics

but is is also wise to consider degradation (of primers, of samples, etc) to explain uneven results

also, when you have two possible products, even if you measure carefully you will not always get the same distribution, right?