library prep problem

[ajthomas]

We have been having trouble getting good libraries for the last few weeks. After doing the library prep, the libraries look just fine on the Bioanalyzer and the concentration is also right where it should be. Then, we do the emPCR and we get no beads at all.

This is true of every library we've made in the last 3-4 weeks. There are 4 that we made first and they had this problem. Another person remade two of them and those to did the same thing. In the mean time, before we figured out it was a library prep problem, he made 24 more libraries. Testing of a few of them showed the same results. We also had a demo of a robot to automate the library prep, and the libraries made by the robot also looked great but didn't work in emPCR. In the meantime, I ran some emulsions with one of these new libraries and an old library that had already been sequenced. The old one worked worked just fine but the new one failed. Finally, these libraries were prepared from library prep kits from two different lots.

We've ordered new library prep kits but they seem to be on backorder.

I don't know for sure what's wrong with these libraries, but my best guess is that the kinase didn't work. If that were the case, the adapters would be ligated on one strand but not the other. This would make them quantify fine and look fine on the Bioanalyzer, but they would fail in emPCR.

Has anyone else had trouble of this sort? Any other ideas what could be the problem and how to solve it?

[RCJK]

I haven't come across this problem (yet). Could you post the lot numbers of the library reagents so others can compare and see if they have used them successfully or had any similar issues?

[ajthomas]

The lot numbers are:
93815820
93803060

[Coyk]

Hi everyone,
I have been looking over old DNA extraction protocols and came across an interesting tidbit. According to Sambrook et al Molecular Cloning 2nd edition, if you use ammonium acetate as the salt to precipitate the DNA, then the ammonium ions can inhibit DNA phosphorylation by T4 polynucleotide kinase downstream. From a little bit of poking around the internet, it seems that in the Qiagen Puregene protocol for DNA extraction, their protein precipitation solution has ammonium acetate in it. I'm not sure if other common commercial kits do as well. Perhaps this is leading to decreased efficacy of the adapter ligation. We've been having inconsistent results with our libraries as well.

[suefo]

Hi everyone,
I have been looking over old DNA extraction protocols and came across an interesting tidbit. According to Sambrook et al Molecular Cloning 2nd edition, if you use ammonium acetate as the salt to precipitate the DNA, then the ammonium ions can inhibit DNA phosphorylation by T4 polynucleotide kinase downstream. From a little bit of poking around the internet, it seems that in the Qiagen Puregene protocol for DNA extraction, their protein precipitation solution has ammonium acetate in it. I'm not sure if other common commercial kits do as well. Perhaps this is leading to decreased efficacy of the adapter ligation. We've been having inconsistent results with our libraries as well.

We are using the AMPure XP kits. Any idea if this contains ammonium acetate? So far all I've seen in a high amount of NaCl and PEG.

[Coyk]

A very rudimentary internet search doesn't yield any evidence for ammonium acetate in the Ampure reagents. I'm not positive though. As a follow-up to my previous post, I also found that Puregene lysis buffer has ammonium chloride in it, which is further evidence that this method probably shouldn't be used for DNA extraction if you want to make 454 libraries.

[pmiguel]

If I recall correctly, the idea of using ammonium acetate, rather than other salts, was that trace amounts of it could be removed from a solution using a speed vap. I have not idea if this is true or not, however.

--
Phillip

[Coyk]

I haven't encountered that in the little bit of researching I've done. The ammonium can't be completely inhibiting reactions or we would've had total failure across the board. The advantage of ammonium acetate according to what I've read is it seems to be the salt of choice if one is particularly worried about rogue dNTPs hanging around.

[ajthomas]

Our problem still hasn't been solved. Here's an update:

We sent one of our libraries off to a friend who runs a sequencing center at another university. It worked just in emPCR for them, so they sent the beads they obtained back to us. In the mean time, we got a new emPCR kit. The emPCR on that library and another one worked just fine with the new kit. We thought perhaps the emPCR kits had a bad component (probably either the enzyme or dNPTs) in them such that they just couldn't amplify the longer molecules in genomic libraries sufficiently but could amplify the shorter molecules of amplicon libraries. That idea seemed to hold up. However, after that one successful run, we went on to try to sequence a couple more libraries and once again the emPCR failed. Then we tried two other libraries that are pools of 12 barcoded libraries. Once again they failed.

Last week we ran a 16-region plate with several amplicon libraries on it. On that plate, we had a little room, so we ran the few beads we got from the emPCR on the barcoded libraries. (We got about 200K beads from a LV prep). Our thought was to see if there was any sequence on those beads or not. If we were to get sequence from those beads, then the library was probably okay, just very low titer, or there was something wrong with the enrichment. We got a total of 5 non-control bead reads from the two libraries. So, there's apparently no DNA on those beads, and those beads are the ones you get from every emPCR that yield no sequence in spite of attaching to the enrichment beads. Now, we're back to thinking that there's something wrong with the libraries, but we don't know what or why. It could still be something with the emPCR, but if so, why would it not be solved with a new kit?

[pmiguel]

Are you always using the same thermal cycler? You might want to run diagnostics on it, if possible. (Some thermal cyclers have this capability.)

Can you look at emulsions under a scope, to make sure your emulsion maker is not malfunctioning. Also, could your libraries have a volatile organic (eg, ethanol) in them that is either damaging the emulsions or hindering PCR? For individual sample sequencing (Sanger) we always speed vap samples to lower the ethanol concentrations of submissions. If your friends sequencing center does this, or some other sort of sample clean up prior to emPCR, that might explain the issues you see.

I should add that when we start running into situations like the one your describe (somewhat intermittent, difficult to diagnose failures) they generally are found to have multiple factors involved. That is any one them does not result in failure, but two or more might. So I would suggest rounding up the usual suspects: pippetters or scales out of calibration, thermal cycler(s) near failure or malfunctioning, temperature/humidity issues, hi-bind plastic ware, etc.

--
Phillip

[ajthomas]

We are always using the same thermocycler. We have considered it, but don't think that's the problem. On more than one occasion, we have run an amplicon and ligated library at the same time (same plate), and the amplicon library will work fine while the ligated library fails.

Any organic contamination would be so low it's hard to imagine it being a factor. Such contamination couldn't be more than 1% at the end of the library prep process, and the libraries are then diluted another 1000-fold or so before adding them to the emPCR, where they get diluted another 200-fold or more.

Thanks for the idea that it could be multiple things. We'll keep that in mind as we do more investigation.

[ajthomas]

One other thing I forgot to mention earlier:

These libraries that aren't working for emPCR will amplify fine in regular PCR. We tried running a standard PCR using a couple of these libraries and an old library we know works using several dilutions. We used 10^7 down to 100 molecules/reaction and the results were the same. We also ran an old library and a couple of the new ones on a digital PCR chip on our Fluidigm Biomark and once again they looked just the same. So, these new libraries that won't amplify in emPCR can amplify just fine in a regular PCR.

[ECO]

What is the final purification method used for your ligated libraries versus your amplicon pools?

[ajthomas]

These are libraries prepared using the standard Roche Rapid Library Preparation protocol. The final purification uses AMPure beads. These are the same beads and ethanol that were used with the earlier libraries that worked just fine. The amplicons have come from others and I don't know how they were purified.

[pmiguel]

Weird. Any possibility that the adapters are not phosphorylated? Rapid ("Y") adapters should be. But if they were not it might explain what you see:

(1) There is a strand denaturation step prior to emPCR for rapid libraries. If the adapters were not phosphorylated, only the "A" strand of the adapter would be covalently attached to the insert So during denaturation you convert your dsDNA molecules into ssDNA single-adapter molecules. Inside the micro reactor, these will not amplify--they lack one of the priming site. However:

(2) In bulk PCR (or in Fluidigm PCR), in cases where the library molecules do not dissociate, the polymerase could add the missing priming site by copying from the other strand. Kind of like what a nick translation does. Except since these are Y-adapters, both ends would then be "A" adapters. So your PCR would be a single primer amplification.

(3) You friend may not have realized these were rapid libraries and failed to denature them prior to emPCR. Then you would get enriched beads because the same faux nick translation could occur inside the microreactor. Alas the beads should not give good sequence, because the two strands of the rapid library amplicon have the A side against alternate sides of the insert. So you would get double (mixed) sequence.

Wait! No, this can't be right. A-A amplicons would not be bound to the bead. So you would not get enriched beads. Well, your friend's result remains unexplained.

A possible test would be to take aliquots from a rapid library that does enrich, and one from a rapid library that will not produce enriched beads. Denature both aliquots (95 oC for 2 minutes, followed immediately by snap chill on ice). Then run them on a pico RNA chip. If you see a large peak at ~40 nt, that would suggest the B strand of your Rapid adapter is not ligating. (There may be some failure to ligate even with successful libraries. Hence the negative control.)

--
Phillip