Make your own AMPure XP, cheaply

[jwfoley]

Like a lot of labs, we've replaced spin/vacuum columns with magnetic bead extraction for every part of our library construction protocols. However, the commercial products are expensive. We found some recipes for making your own SPRI bead mix at a tiny fraction of the cost, but they weren't written very well, so here's a new one.

Enjoy!

---

UPDATED April 17, 2015: we now have a protocol for RNA beads (comparable to RNAClean XP) as well

UPDATED AGAIN October 17, 2015: there was an error in the RNA bead mix volumes

UPDATED AGAIN October 21, 2015: there was also an error in one of the NaCl concentrations (thanks, blameyourelf!)

FINAL UPDATE February 16, 2015: we've moved the protocol to OpenWetWare for easier updating; please go there for the most recent version

[Simone78]

Nice, thanks for sharing!
I guess there are a lot of similar protocols out there. I am using the one described in Rohland and Reich, Genome. Res. 2012, 22: 939-946 and it works perfectly. Adding more or less PEG 8000 to the final buffer will change the cut-off of fragments that are recovered (more PEG > more short(er) fragments and viceversa).
No more AMPure beads anymore!

[jwfoley]

Have you found that Rohland and Reich's recipe works at low volume ratios (0.7X or 0.6X)? It has a slightly lower PEG concentration and especially a much lower NaCl concentration, so I'd worry that it might lose the ability to salt out the DNA efficiently. Ours works down to 0.5X, though that's probably lower than anyone would realistically want to go for the current short-read sequencing platforms.

But the main reason we wrote out a longer protocol is that there are some non-obvious problems we solved just for mixing up the solutions (Tween 20 is too viscous to measure by volume and PEG 8000 is difficult to measure and dissolve). If you don't write these things out, people do all sorts of crazy irreproducible things, like measuring liquids in conical tubes (one protocol even specifically says to do that!).

At any rate, yes, the PEG is the ingredient that determines size selection. Another thing that has worked in our hands (the one time we tried it) is preparing a low-PEG SPRI mix to clean up DNA from a reaction that already has PEG in it (e.g. ligation). If you know from previous testing what ratio gives you the desired size range from a "clean" sample, you can adjust the PEG in your SPRI mix so that the final concentration in the reaction + SPRI volume is the same as that, and it gives the same result. If you were working with AMPure, you might have to do a large dilution of your reaction and then use a small ratio of bead mix to compensate, and then we're back to worrying about the salt. Of course you should ideally calibrate your ratio with your actual working solutions anyway.

[Simone78]

Have you found that Rohland and Reich's recipe works at low volume ratios (0.7X or 0.6X)? It has a slightly lower PEG concentration and especially a much lower NaCl concentration, so I'd worry that it might lose the ability to salt out the DNA efficiently. Ours works down to 0.5X, though that's probably lower than anyone would realistically want to go for the current short-read sequencing platforms.

I use my own tagmentation protocol recently published (Picelli et al., Genome Res 2014). In the last step I clean up with 0.8X-1.0X beads (24% PEG in the buffer but otherwise exactly the same as R&R). The 24% is the % that allows the recovery of 200-300 bp fragments (in our hands, at least), yet keeping the primer dimers to a minimum (which actually could be achieved in a smarter way, just decreasing the primers in the final PCR, as we say in the paper). Anyway, 0.8X works well, below that we saw that we were losing PCR product more and more.

[Andy Flavell]

Cost gone up?

Hi
I've been trying to track down these beads and they are coming up at a much much higher price than you mention - e.g. GEC Catalog No.: 09-981-122 100ml $1647. They are "GE Healthcare Sera-Mag™ Magnetic SpeedBeads™ Carboxylate-Modified; Dia.: 1µm; 7 EDAC/PA5; 100mL"

Maybe they've just plugged a loophole?

Andy

[jwfoley]

No, that sounds right; 100 mL of the Sera-Mag beads will make 5 L of SPRI mix, and that much AMPure XP would cost about $72,000 CAD at our price (though we could probably get a volume discount). Are you sure you need that much? There's also a 15 mL size available for the Sera-Mag, using the part number in our protocol.

If it helps, the beads are also resold by Fisher and Sigma-Aldrich.

[Andy Flavell]

Brain fade - I forgot the 50-fold dilution step!

Thanks for the help

Andy

[kerplunk412]

Thanks for sharing, I just looked over the protocol briefly and it looks great!

One thing I would like to add if you don't mind, I have found that mixing sample with about 50 ul of AmpureXP and then adding an equal volume of isopropanol (so ~50 % IPA final) can be used to purify pretty much all sizes of nucleic acids very efficiently. The idea is based on a protocol that Beckman released for purifying miRNAs with SPRI. I am happy to say I have not performed an ethanol precipitation or column clean-up for DNA or RNA since I discovered this.

[Manna]

I ordered these SpeedBeads Jan 29, and they haven't shown up yet (Feb 9), so I called Fisher. Their rep said they were a custom item, and it may take 1-3 months before I actually get them. Is this what others have experienced?

[jwfoley]

Yes, actually, we had the same problem but I thought it was just a temporary stock issue. Some of the websites will actually tell you whether they're in stock before you order.

At least once you get them, you're set for a while (15 mL of Sera-Mag SpeedBeads yields 750 mL of SPRI mix).

EDIT: I should add that we had this experience ordering with Fisher in Canada. Has anyone had better luck with Sigma-Aldrich or another reseller?

[Manna]

OK, thanks. Looks like I'm in for one more bottle of AMPure.
Now... to tell the boss or not to tell him?

[thermophile]

I've also used Ampur but made my own PEG/NaCl, so I could clean 100ul pcr product with just 20ul Ampur. I was never saturating the beads with my pcr product. I did have to recalibrate the PEG every month or so-no obvious precipitation but the size selection would shift towards larger fragments if the PEG was more than a few weeks old (i.e. instead of adding 30 extra ul PEG to the 20ul beads and 100ul PCR product, I'd have to add 35 or 40ul)

[jwfoley]

We now have a protocol for RNA beads (comparable to RNAClean XP) as well. I've added it to the original post.

[jcaner]

Have you found that Rohland and Reich's recipe works at low volume ratios (0.7X or 0.6X)? It has a slightly lower PEG concentration and especially a much lower NaCl concentration, so I'd worry that it might lose the ability to salt out the DNA efficiently. Ours works down to 0.5X, though that's probably lower than anyone would realistically want to go for the current short-read sequencing platforms.

But the main reason we wrote out a longer protocol is that there are some non-obvious problems we solved just for mixing up the solutions (Tween 20 is too viscous to measure by volume and PEG 8000 is difficult to measure and dissolve). If you don't write these things out, people do all sorts of crazy irreproducible things, like measuring liquids in conical tubes (one protocol even specifically says to do that!).

At any rate, yes, the PEG is the ingredient that determines size selection. Another thing that has worked in our hands (the one time we tried it) is preparing a low-PEG SPRI mix to clean up DNA from a reaction that already has PEG in it (e.g. ligation). If you know from previous testing what ratio gives you the desired size range from a "clean" sample, you can adjust the PEG in your SPRI mix so that the final concentration in the reaction + SPRI volume is the same as that, and it gives the same result. If you were working with AMPure, you might have to do a large dilution of your reaction and then use a small ratio of bead mix to compensate, and then we're back to worrying about the salt. Of course you should ideally calibrate your ratio with your actual working solutions anyway.

Hi! I'm try to clean a double digestion with your protocol but the DNA recovery after the cleaning is very low... We could only recover arround 20% of our DNA. This is the bead concentrations we tried:

1.5X: 49.8 ng/uL (before), 10.4 ng/uL (after) = 20.88% efficiency
2x: 54.8 ng/uL (before), 10.7 ng/uL (after) = 19.52% efficiency

I had followed all the steps very carefully. Doing all the vortex, pipetting and mesures as accurate as possible.

I really don't know why I only can recover that few amount of DNA. So, could you please have any advise or known what could have happened?


Thank you in advance!
J.

[luolipang]

Does anyone notice that the buffer is getting viscous as long as the tween was added？While the AMPure XP is not that.