Using Lower Input DNA Concentration for Nextera XT

[MicroBio]

Hi all,

I have VERY low amounts of viral DNA which I would like to sequence using the MiSeq platform. The DNA concentrations are all below the 0.2 ng/uL recommended by Illumina (they suggest using 1 ng total). Some are not even detectable by Qubit.

Does anybody have experience using concentrations of DNA lower than recommended for Nextera XT? I was thinking that increasing the tagementation incubation time might increase the amount of DNA amplified by the PCR step, and therefore increase the end yield. I was also thinking about increasing the number of PCR cycles from the recommended 12. I thought it was worth a try to see if anybody had advice or suggestions. I am still somewhat new to this.

Thank you!

[koadman]

We have been using lower input quantities with the Nextera XT kits, although it requires some fine-tuning to get the normalization step right. As discussed in an earlier thread in this forum, it seems like the main issue is that with different input quantities, the resulting fragment size distribution can be different in different samples, and this changes the total number of template molecules required to saturate the normalization beads. As for the number of PCR cycles, we upped it to 14 but qPCR prior to normalization on individual samples with 500pg input suggests that we have far more template than needed for a flowcell (e.g. 50x more). So there may be no need to add PCR cycles.

[Bucky]

We also use the kit with lower input concentrations (~ 0.1 ng total or less). We are using 15 cycles of amplification and instead of their recommended normalization we use qubit and bioanalyzer to determine the number of DNA molecules. Works really well for us. You can tweak the Ampure XP clean up a bit in order to get larger fragment sizes (our overall fragment length will be shorter due to less input material).

[ScottC]

Also, if you're in a hurry and have throughput to waste (i.e. not that many viral genomes per run, giving you far more coverage than you need) then you can just top up your sample DNA with something known. That way, you get the coverage you need and you don't have to do any optimisation. Just prep and run. We've used commercial phage Lambda DNA before for this purpose.

[MicroBio]

Thanks for the help everybody. Bumping the number of PCR cycles to 15 from 12 worked great. We have not sequenced the libraries yet, but right now they are looking preliminarily good for samples with 0.1 ng/uL or less.

[oolong]

Nextera XT prep on PCR amplicons

I'm running into trouble preparing the Nextera XT library from (15 cycle) inverse PCR amplicons (500bp to 1200bp size). I know there's product in there since I checked all of them with a 14-cycle nested PCR reaction. The sample concentrations for my inverse PCR products are at slightly less than 0.2ng/ul (Qubit HS assay), and the core lab is not comfortable moving forward with the Nextera XT prep if they can't detect amplicons on the BioAnalyzer (supposed 5pg/ul). In fact, we've already done this once, and the final PAL library was qc'd at 27.1pM. I read this forum before and bumped up the Nextera XT PCR cycle to 15, yet the MiSeq run failed when the PAL was cut 25x as per Illumina loading instruction.

My question is, should I simply use the nested PCR products to do my Nextera XT prep since ? But I'm worried about GC bias affecting the sequencing readout. I've called Illumina many times and each time, but they won't give a definite range for the PAL concentration needed for optimal cluster formation. I am multiplexing my samples but am not concerned about balancing the library since I'm only after identification of the invPCR products. The ordeal is taking over 2 months and there's no one here with experience using the Nextera XT so I'm asking for advisement from more experienced users.

Many thanks!

[kcchan]

Have you tried not doing the normalization step at all and instead pooling it normal as you would with a TruSeq library? I find that this method works much better with samples of low concentration.

[MicroBio]

Hi oolong. We always do BioAnalyzer on the CAN samples (not the PAL) right after the PCR cleanup step. We do this because after library normalization the pool is ssDNA which will not work well on the BioAnalyzer (according to the manual). Could this be why the core is not getting anything on the BioA trace?

[oolong]

This time, the BioAnalyzer was run on the input PCR amplicons (pre-tagmentation). I didn't run the Bioanalyzer post-tagmentation step the previous time. I've decided to go ahead and do a nested PCR on all my samples before attempting the Nextera XT kit again. Skipping the normalization beads step sounds reasonable since we aren't expecting many unique sequences (~2000 max). If my library is between 250-500bp size, should I run 2x250 , or is that overkill?

[MicroBio]

Hi oolong. This might be a little late and you might have already decided, but I'll chime in anyways. I would think that, if you wanted to have more paired end sequences that overlap in the middle, then it might be good to do 2x250, but other than that I'm not sure if there are any other benefits. Maybe somebody else on here has more insight to offer?

[kternus]

Hi everyone! My lab is interested in using 15 PCR cycles with Nextera XT for low DNA concentrations, and I was wondering if anyone had published their use of that method?

[GSviral]

Hi kternus,

I use the Nextera XT library prep kit with 15 cycles of PCR. Originally, it was for low concentration viral DNA/cDNA and I found it helped boost the yield after the index PCR to get greater success for sequencing. This was a metagenomics workflow however so I wasn't as concerned about PCR duplicates.

Additionally, I use 15 cycles of PCR when performing whole genome sequencing of bacterial genomes with a standard 1 ng total input. Again, after some testing, the PCR duplicates haven't had an effect on the analysed results and I am getting really high coverage across the 98% of the genome. The 15 cycles in this situation is probably overkill but I have gotten the preps and sequencing to such a consistent level of yield and clustering that I don't want to alter it any further!

Hope this helps!

[UCan'tBcereus]

Hi kternus,

We have also also worked on some very low input Netera XT samples as well, although we were working on metagenomic samples too.

We used the following paper from the Phillip Hugenholtz lab as a general guideline:

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5036114/

They use 20 cycles of PCR and dilute the transposase 1:10 to help prevent the transposase from fragmenting the limited amounts of DNA too much. This paper discusses PCR duplicates and contamination issues as well.

We ended up having to do two Library purifications (both 1X cleanups) to get rid of all the primer dimer, but it did work. Most of these DNA samples were below detection on the Qubit dsDNA HS kit (i.e. below 0.05 ng/uL) before library preparation. We also ran a negative control as well.

Generally if I got a reading of 0.05 ng/uL or above on the DNA, I could make a successful Nextera XT Library with the standard 12 cycles (libraries ending up between 0.3 ng/uL to 2 ng/uL, and correctly sized.

Good Luck!

[kternus]

Hi GSviral and UCan'tBcereus (great name! ),

Thank you so much for sharing your feedback! That's really helpful, and I'm happy to hear that you've had so much success with increasing the number of PCR cycles. I appreciate the additional technical information about your previous experiments, which will definitely benefit my team in our upcoming sequencing project.

It's really useful to have a published citation to reference, so thanks for sharing that publication from Philip Hugenholtz's lab!