Nextera Kit -- DNA fragments too small

[arilindsey]

Hello --

I have been working with the Nextera kit for Illumina to create DNA libraries for genome sequencing. I initially used the recommended 50ng input DNA (quantified with a Qubit). However, what should have been my final library had no DNA, and instead all the DNA was left in the supernatant that comes off the first Ampure XP bead clean up step. When I run that same supernatant on a gel, the DNA fragments range from 80 - 200 bp. I have tried higher concentrations of input DNA (up to 90ng) and still have the same problem. Not sure what to try next -- has anyone experienced this? My input DNA has a high molecular weight -- 20kb+ fragments.

Any help is much appreciated.
Cheers,
Amelia

[Jfly7]

Hi Amelia,
I found this question and I am wondering if you had any luck making adjustments to make longer libraries with Nextera? I prepared libraries with Nextera last year and also got shorter fragment sizes than ideal (200-250 bp) when I would like 300-400 bp. I am wondering if it would be best to use a higher concentration of DNA or lower concentration of enzyme?

[NextGenSeq]

Single cell RNA-Seq libraries with an input of 1 ng or less of cDNA give Nextera fragment sizes of over 500 bp.

Note with the HiSeq 3000/4000 all libraries have to be less than 500 bp so you might want to reconsider whether you want larger fragments.

[Jfly7]

Thanks for the reply.
I am doing genomic DNA library prep and am using HiSeq 2500 100bp PE reads. The reason I want libraries a bit longer than what I got is because a lot of fragments were short and therefore sequenced into the adapter, so we lost a fair amount of sequencing coverage. I would like to avoid this this time around.

[NextGenSeq]

What I find most annoying about Nextera is the insertion bias.
Every Nextera I've ever seen shows unusual GC bias in the first 15 basepairs (see attached).

[pmiguel]

Originally Posted by Jfly7

Thanks for the reply.
I am doing genomic DNA library prep and am using HiSeq 2500 100bp PE reads. The reason I want libraries a bit longer than what I got is because a lot of fragments were short and therefore sequenced into the adapter, so we lost a fair amount of sequencing coverage. I would like to avoid this this time around.

Just use less Ampure during the clean up.
You will get a wide range of sizes using Nextera. Getting the insert sizes of amplicons that are sequenced to increase in length is just a matter of removing the shorter amplicons. (Shorter amplicons seem to cluster preferentially.)

--
Phillip

[pmiguel]

Originally Posted by NextGenSeq

What I find most annoying about Nextera is the insertion bias.
Every Nextera I've ever seen shows unusual GC bias in the first 15 basepairs (see attached).

Nextera is Tn5. When you recruit a natural protein to do a job for you, you recruit its natural proclivities as well. Sure they have engineered it a little, but not enough to completely remove its insertion site preferences.

--
Phillip

[nucacidhunter]

An interesting patent by Agilent was recently published about role of bivalent cations on Nextera activity.
http://patft1.uspto.gov/netacgi/nph-...S=PN/9,005,935

[pmiguel]

Originally Posted by nucacidhunter

An interesting patent by Agilent was recently published about role of bivalent cations on Nextera activity.
http://patft1.uspto.gov/netacgi/nph-...S=PN/9,005,935

Very interesting!
Does Agilent actually sell this Vibhar Tn kit?
So it is likely that including Mn in Nextera reactions would reduce the site specificity. Interesting...

--
Phillip

[Simone78]

Interesting! When using our home-made Tn5 the tagmentation was actually INHIBITED by MnCl2 and worked only with MgCl2 (with 5 mM final being the optimal conc, while higher conc had a deleterious effect).
Thanks for sharing,
Simone

[nucacidhunter]

Originally Posted by pmiguel

Does Agilent actually sell this Vibhar Tn kit?

--
Phillip

They offer two transposase based kits one for SureSelect QXT capture and another for WGS library prep. There is no info in manuals on transposase used, but for sequencing one must use custom reads and index sequencing primers and use “CTGTCTCTTGATCACA” for adapter trimming which could be full or partial recognition sequence of transposase. Nucleotide BLAST for adapter sequence brings up lots of hits.

[pmiguel]

Originally Posted by nucacidhunter

They offer two transposase based kits one for SureSelect QXT capture and another for WGS library prep. There is no info in manuals on transposase used, but for sequencing one must use custom reads and index sequencing primers and use “CTGTCTCTTGATCACA” for adapter trimming which could be full or partial recognition sequence of transposase. Nucleotide BLAST for adapter sequence brings up lots of hits.

From the text of the patent application:

Using bioinformatics means, the IRL sequence for the transposase was identified as 5'-ctgtctcttatacacaat-3' (SEQ ID NO:9) and the IRR sequence as 5'-acttgtgatcaagagacag-3' (SEQ ID NO:4). Through experimentation, a modified 19 bp IRR sequence 5'-agatgtgatcaagagacag-3' (SEQ ID NO:5) proved to be more efficient and was used in most experiments.

So the sequence you provide is consistent with Agilent using the Vibrio harveyi IS4-like element.

But, I guess the question still remains -- how non-biased is its insertion? It could be that by adding Mn ions the insertion specificity became lower. It certainly looks that way from the figures in the patent. But it may be just as high as Tn5 in Nextera.

--
Phillip

[nucacidhunter]

I have not seen any data comparing insertion bias to sheared DNA or Nextera. But comparing couple of Nextera and QXT FastQC plots (attached image) shows that it also has insertion bias but different from Nextera.
Nextera vs QXT.pdf

[SNPsaurus]

Thanks for the info, nucacidhunter. I am curious about the QXT kmer plot, though. Looking at the base% plot, the bias is GC at the first 2 nucs then mildly T. How come there are no kmers with that at the start? I guess the kmers shown have some representation across the read. I like to compare kmers at the start to kmers selected from the middle of the read to see differences after normalizing the kmer content of the genome. Pretty interesting, though!

[NextGenSeq]

Originally Posted by nucacidhunter

I have not seen any data comparing insertion bias to sheared DNA or Nextera. But comparing couple of Nextera and QXT FastQC plots (attached image) shows that it also has insertion bias but different from Nextera.
Attachment 3809

That looks like unbalanced barcodes to me.

[DRYTCYV]

Hi,

Anyone tried KAPA HyperPlus Kits?
They say that this is the direct nextera competitor.

Cheers,

[nucacidhunter]

It could be considered a competitor to NEB dsDNA Fragmentase but not Nextera. Nextera does not require end repair and adapter ligation unlike Hyper.

[SNPsaurus]

Our local facility looked at the Hyper kit and it looked less biased than Nextera in terms of k-mer preference of the ends.

I think it is a competitor to Nextera since it approaches it in terms of workflow speed and amount of hands-on time needed. At least I am evaluating it as a replacement for our Nextera-based libraries.

[SCIL2019]

Hi all,

I am having a bit of an issue with this library/sample. I am getting larger fragments peak and then 200 bp peak is gone. This is using 10x ATAC kit.

I usually don;t see such a nhigfh large peak yield, likely because I remove it with dual sided selection. here, I also did dual sided selection but for some reason it's still there...and what's more odd is that I can see the 75 bp dimer too.

Ideas?

Thank you!

[nucacidhunter]

Regardless of size distribution, library does not seem to have typical periodicity expected in ATAC-seq libraries indicating that there was ambient DNA in nuclei stock or nuclei disintegrated during transposition. Presence of 74bp fragment also raises question about accuracy of size selection.