Amplicon sequencing - gel cut purification

[JKistler]

Hi all,

For our next 454 amplicon sequencing run we are thinking to do a gel cut purification of our amplicons in order to eliminate small fragments e.g primer dimers (they may be getting through our libraries unnoticed as we have quite a large percentage of reads failing on 'short quality').

Can anyone recommend an effective kit/method for this which has given good results? Also, as I will pool multiple samples together, could I do this gel cut on the entire pooled library or would it need to be the individual amplicons (it is a 16S library so all amplicons should be the same size).

Any advice would be hugely appreciated!

Many thanks,
James

[ajthomas]

I imagine any normal gel purification method should work just fine. One caution I would offer, however, is to be sure you use SYBR Safe and a blue light box rather than a UV light box. Using UV for the gel cut will cause a lot of damage that may severely impact your sequencing results. I imagine you already are as it seems that many labs have already switched over, so this advice is just in case you haven't.

You could do it either before or after pooling your samples. The advantage of doing it after pooling is that it is much less work, as you state. The only problem that might arise is if you're trying to pool your samples stoichiometrically and you get a lot more primer dimers from some samples than others. In that case, samples that produce a lot of primer dimers will end up underrepresented in the final pool because the primer dimers will make up a larger fraction of the total from those samples before pooling. So, as long as you get similar results from all of your samples I would say do it after pooling.

[mohdsos]

hi JKistler,

you can quantify the amplicons and pool them with consideration of a normalized concentration of all, then you can purify then using ampure XP beads or gel cut.



mo

[pmiguel]

Non-denaturing gel cuts will not get rid of primer dimer strands annealed to full length amplicon strands. And the primer dimer strands will anneal to the full length strands, right? They share adapter sequence.

Roche recommends "double AmPure" at reduced AmPure concentrations. Same problem, I guess. But some of the primer dimers probably do exchange off the full length amplicons during the process. So each added cycle of size selection probably helps.

Might also want to try tuning the PCR reactions to minimize primer dimer formation. Lower primer concentrations, etc.

--
Phillip

[JKistler]

Thanks to all for your advice. I'll go with a gel cut after pooling and optimise PCR as best as possible to reduce primer dimers. Any improvement, would be good. That's worrying to hear that the primer dimers can anneal back to full length strands though.

[ajthomas]

In theory they can, but I haven't personally seen much of a problem from it. If you're using a gel cut purification method, you should get rid of those hybrid molecules as well because they will migrate differently from a regular amplicon duplex.

I wonder if you could get rid of those hybrid duplexes by treating the library with mung bean or S1 nuclease before doing the AMPure purification. That should destroy the single stranded parts of those molecules and leave your double stranded amplicons alone. I suppose that's worth a try for anyone having trouble with short reads that might be caused by such hybrid duplexes.

[pmiguel]

In theory they can, but I haven't personally seen much of a problem from it. If you're using a gel cut purification method, you should get rid of those hybrid molecules as well because they will migrate differently from a regular amplicon duplex.

A primer dimer is pretty short. I was assuming that the amplicon ends up double stranded throughout its length. It just has the primer dimer riding one adapter end, with the rest of the complement-to-the-adapter strand peeled back to allow this. This would only increase the molecular weight of the duplex by the molecular weight of a single strand of the primer dimer. Not enough to remove using size selection unless your cut is very narrow.

Unless you had a pretty bad primer dimer issue, the concentration of these hitchhikers would be low enough to ignore, though.
--
Phillip

[eladSeq]

Hi,
We are having trouble sequencing our amplicon pool due to short read failure causing very low pass filter %. For this library, I combined amplicons and then did an agarose gel extraction using a UV gel box. Then followed this with AMPure XP purification.

We have successfully sequenced a similar amplicon pool this way, so I didn't think the UV would cause a problem. The only difference between this library and the previous successful one was pooling more amplicons with different barcodes, so I guess the addition of these new primers could be causing more primer-dimers? On the fragment analyzer we only see a single clean band at 421 bp, but I guess the primer-dimers could be hiding in there by hybridizing to the full length amplicons?

Have you guys had success eliminating the short read failures by reducing the primer-dimers, or eliminating the UV gel extraction step? Any suggestions or advice appreciated. Thanks!