[MissDNA]

Our case was the emPCR kit for sure because the library we were trying to amplify had been sequenced before in our lab. We have 14 boxes of emPCR reagent that will expire now in Nov. we will have to use them. The reagent that expire first is only the capture wash buffer. I don´t think we will have much trouble because of it.

[pmiguel]

That's usually true, but not always. We had an issue a few months ago where we couldn't make a functional library for several months. The libraries were of a good concentration and looked beautiful on the BioAnalyzer, but they always failed completely in emPCR. We had the same issue with a few different library prep kits, all from the same lot. It wasn't until we switched out the enzymes in the kit with enzymes from another manufacturer that it worked. We then got replacement library prep kits from Roche and everything was fine again. I don't know for sure, but I think the problem was probably the PNK. If that is bad, then the fragments won't be phosphorylated, and the adapters will be ligated on, but only on one strand. The result will be fragments that will quantify just fine because the fluorescent tag is attached, but once the duplex is melted there will be only one adapter on each molecule and the emPCR will fail.

Were these rapid libraries? If so, did you assay them using fluorimetry?

We would have picked up an issue of this sort via qPCR, I think.

--
Phillip

[pmiguel]

Another question: you talked about blunt and A-tailed adapter. I didn´t actually know there was a difference. Is that difference due to the fact one library is single stranded and the other double?

No, for rapid libraries the adapters have a 3' T overhang. This largely prevents adapter-adapter ligation. The inserts must be A tailed to be compatible with the adapters. Taq polymerase, for example, will add a single non-templated 3' A to a blunt dsDNA fragment end. This also decreases insert-insert ligations (chimeras).

--
Phillip

[ajthomas]

Were these rapid libraries? If so, did you assay them using fluorimetry?

We would have picked up an issue of this sort via qPCR, I think.

--
Phillip

Yes, these were rapid libraries. We did quantify them using fluorometry, not qPCR. We don't have the reagents for doing it that way. We did try a regular PCR to see if it would amplify and it did just fine. I have some ideas to explain that, but I don't know for sure what was going on there. We never tried replacing one enzyme at a time to nail down the problem for sure. It started working again, so we were relieved.

[MissDNA]

No, for rapid libraries the adapters have a 3' T overhang. This largely prevents adapter-adapter ligation. The inserts must be A tailed to be compatible with the adapters. Taq polymerase, for example, will add a single non-templated 3' A to a blunt dsDNA fragment end. This also decreases insert-insert ligations (chimeras).

--
Phillip

Got it, thanks for the explanation. That explains the use of T4 polymerase in the fragment end repair reaction.

[Dun]

Do you guys all use 454 original paired-end protocol for library prep? It seems to be adapted from general library prep protocol, which is...lengthy.
Anybody tried Bruce Roe lab's updated protocol? http://www.genome.ou.edu/proto.html
Basically, it opts out procedures such as immobilization, amplification and single-strand making. More like a rapid prep protocol.

But as post#3 mentioned,this simplified version has one innate disadvantage:
the library contains "pair-end" fragment and regular shotgun fragment.
According to Roe lab, 3~5% reads are paired-end, the rest are regular shotgun.

Whether is it really worth doing a paired-end run with 95~97% of the reads are regular shotgun? I am still in the preparation stage.
Can anyone with experience on paired-end assembly give me an idea? THX

Dun

[pmiguel]

We dropped the single-strand stuff long ago. Too failure-prone in our hands. Also, why bother -- PCR suppression effect should remove the majority of the A-A and B-B constructs.

But otherwise we do use the Roche technique. Even used it to construct an Illumina mate end library.

I remember the Roe lab technique pre-dating the Roche technique.

5% paired end reads might be enough. BTW, if you use this method make sure not to add that exogenous DNA that the Roche protocol has you add prior to the second shearing. We don't add that DNA anyway -- crazy! But it would be really bad to add it and not do the immobilization step, I think.

--
Phillip

[pmiguel]

Got it, thanks for the explanation. That explains the use of T4 polymerase in the fragment end repair reaction.

You mean for the old (standard) protocol? Yes, I guess. T4 pol will both fill-in 5' overhangs and degrade 3' overhangs. The T4 PNK will add phosphates to 5' ends and remove them from 3' ends. These are very frequently used to create blunt, but correctly phosphorylated DNA ends after non-enzymatic shearing methods.

In the rapid protocol, the Taq probably does not do much during the 20 minute incubation at 25 oC. Then the T4 pol will be inactive during 72 oC step during which Taq will tend to add non-templated single base 3' A-tails on the otherwise blunt insert fragments.

Hmm, I have not looked at that protocol for a while. If you wanted to ligate rapid adapters instead of standard adapters, I would suggest using 3'-5' exo- klenow to A-tail, rather than Taq. Probably could do the reaction on the immobilized DNA if you wanted, then use an MPC purification method to remove it prior to adding the rapid adapters and ligase. Obviously, you would skip the fill-in reaction (the rapid adapters are phosphorylated -- no need.)

Anyway, not completely trivial. Approach with caution, etc.

--
Phillip

[Dun]

THX Phillip

Yes, adding some "carrier DNA" sounds very weird to me as well.
Seems Roche is confident about removing it completely during immobilization, which is unlikely.

Roe lab protocol has newer (2009 Oct) version, that is after Roche's current one.

We dropped the single-strand stuff long ago. Too failure-prone in our hands. Also, why bother -- PCR suppression effect should remove the majority of the A-A and B-B constructs.

But otherwise we do use the Roche technique. Even used it to construct an Illumina mate end library.

I remember the Roe lab technique pre-dating the Roche technique.

5% paired end reads might be enough. BTW, if you use this method make sure not to add that exogenous DNA that the Roche protocol has you add prior to the second shearing. We don't add that DNA anyway -- crazy! But it would be really bad to add it and not do the immobilization step, I think.

--
Phillip

[MissDNA]

Thanks, Phillipe. I do not intend to modify anything in the RL protocol as it works well for us. As for the PE protocol, I have only tried the 3kb one and the only modification we did was remove the first ampure after shreading and use a QIAquick column instead. Usually the yield after nebulization is very low and I tend to follow with 2 library amplification a reaction and combine them after the melt step in the same MinElute column.

The protocol is a bit tedious, and the fastest I could do it was in 2 days, which is not that bad considering other molecular biology protocols. I have to construct 5 libraries in the next few weeks, and since I actually will have to supervise the other lab techs doing it, I am not looking forward to. I rather do it myself but they need to learn. I am yet to try 8 and 20kb, which I do believe are way more challenging and complicated. That gel/elutrap step for size selection really bugs me.

I asked all bioinformation about the rate of true PE we get, and he said sometimes about 70%. He did tell me that it could be nice to see what is the minimum amount of true PE needed to generate scaffolds, and if the rest was all SG reads would be actually good to increase coverage.

[MissDNA]

About the carrier DNA: what its role, anyway?

The first PE prep we tried was the old LPE protocol for FLX standard chemistry, in that protocol some E. coli DNA was used as carrier and our first run turn out to be all E. coli!

[Dun]

Carrier DNA might be used as "carrier" in the follow procedures, I guess. Because the input DNA is pretty low (100ng), might need material to carry on.

[pmiguel]

I always presumed the extra DNA was to facilitate shearing. I don't think nebulization works well at low concentrations of DNA. (Use sonication, if available...)

I have probably mentioned this before, but several times I have seen sequencing projects run afoul of people using yeast tRNAs as co-precipitants. It just seems perverse to take a sample you went to some effort to purify and then deliberately add a contaminant to it.

--
Phillip

[MissDNA]

Long time ago, when we had the problem with PUC19 being used as carrier, some Roche tech support told us we could spike the sample with some gDNA from the same organism. I forgot the details but it makes sense now. Also, I think the current DNA carrier does not amplify in the PCR or something like that, so in the end it would not contaminate the final library.

[pmiguel]

The PE protocol includes an immobilization step that relies on the biotinylated circularization adapter binding to the streptavidin magnetic beads. So, in principle, only molecules ligated to the adapter should be retained.

So, in principle, you can through any DNA you want in at that stage and little of it should show up in your library. In principle.

--
Phillip