Avoiding primer dimers

[lpalacios]

Hello,
I am working with the 454 but I have problems with the PCR multiplex because I don't manage to avoid primer dimers. I carried out 2 consecutive PCR and the dimers reach to about 115-130 bp so they are not removed by Ampure purification. Do you have any other purification method that eliminate dimers or small inespecific amplicons of these sizes? or any recommendation to solve this problem?
Thanks in advanced,

[clarancer]

When I see primer dimers I change the ratio of AMPure XP beads to PCR. Rather than add 72ul of AMPure beads to your 22.5ul PCR and 22.5ul water add 45ul of beads. Follow the protocol for the remainder and you should be fine- works for me every time. Your yield may be slightly lower but at least it will primer dimer free!

[lpalacios]

Thanks for your answer clarancer, but I have just tried it and the dimers don't disappear since their lenght is too long...

[Palecomic]

We have lots of primer dimer, and it is tricky to remove, but have had success with an initial ethanol/sodium acetate precipitation (to pool replicate PCRs together) followed by two rounds of Ampure XP purification (the first round I resuspend in 22.5 µl to make it easy for the second round), with an altered ratio of 0.7 beads to amplicon (31.5 µl of beads to 22.5 of amplicon and 22.5 of water).

That should remove dimer beneath 200 bp I think, but the two rounds are definitely required. In addition our primers are shorter, around 70 bp and our amplicons long (around 670 bp), so we can size select quite harshly.

This ratio is a bit harsher than clarancer's so may be better in your case, but it's the ratio that you need to play with, the less beads, the less effective it is at precipitating the shorter sequences. Bear in mind that you will start to lose amplicon if you go to far.

[TonyBrooks]

It's not a very high throughput method, but gel-cut works for us.

[OnUrMark]

It's not a very high throughput method, but gel-cut works for us.

Hi Tony, would you mind sharing your gel extraction technique? I believe there may be contamination from our final gel extraction that is inhibiting the sequencing reaction. The bioanalyzer shows there are no primer-dimers or adapter-adapters present in our final product.

I originally posted on this problem here: http://seqanswers.com/forums/showthread.php?t=10399

Thanks so much for your help!

[TonyBrooks]

We just make a 2% BioRad Low Range Ultra Agarose gel with 1X TAE (50X from Sigma), adding 2µL of SafeView Stain (NBS) to every 100mL of gel before pouring. Sizing is done with 2µL of 1µg/µL 1kb Plus ladder (Invitrogen). Ideally, run a ladder on a few wells to check for uniform migration of fragments. Leave at least two empty lanes between the ladder and any sequencing sample.
Don't run too many samples on the same gel, we run a maximum of 2, leaving 8 lanes between samples to avoid cross contamination. Run at 150v for 90 mins in 1X TAE.

Image the gel under UV and work quickly to limit exposure. Use a face guard (remember health and safety!)
Excise the each library using a disposable scalpel and transfer into a 15mL Falcon tube. We usually cut around 100bp out from each sample between 250-350 bp. Samples are weighed on an electronic scale, zeroed on an empty 15mL Falcon. Weights are typically 400mg. We then follow the Qiagen QiaQuick Gel Extraction protocol, including the isopropanol step, eluting in 30µL of EB.

[OnUrMark]

Hi Tony, Thanks for sharing your gel extraction protocol. I have not been leaving more than one lane between samples and sometime the ladder is in the adjacent lane. I am surprised by the high voltage, for some reason I was under the impression my gel would melt.

I am multiplexing 6 samples per lane, and it was recommended to me that I run all samples for a lane on the same gel. Can someone tell me if you have noticed that this is important or not important? Thanks.

[HESmith]

Running multiplexed samples in the same lane for gel extraction insures that all of the samples are the same size range (it's nearly impossible to cut the identical fragment size from multiple lanes). However, there's no way to adjust the relative concentrations of the different samples from that point. In my experience, the ability to normalize the concentrations prior to sequencing more than outweighs having the identical fragment size (although that might depend upon your particular application).

[TonyBrooks]

I guess you could run all six on the same gel, cut out each band separately and do 6 gel extractions. You could then quantify and pool before sequencing. There might be some minimal cross contamination but that shouldn't matter too much if you have factored in enough redundancy in your multiplex run.