Comparison fragmentation through Covaris / Tagmentation

[hbn]

Hi guys,

I am looking into the new Illumina library preparation kit protocols (amongst which are the Nextera kits) and I was curious what your experiences are with tagmentation as opposed to fragmentation using e.g. the Covaris (previous protocols)?
Is the fragmentation using tagmentase as random as with mechanical fragmentation? Does is result in more or less even coverage over the genome?

I could not find any real personal experiences with about this just yet..
Thanks for any input!

[Simone78]

Hi,
in the original paper from Adey & Shendure (Genome Biol, 2010) they looked at the bias of Covaris-transposase-DNA fragmentase (NEB). Although there is a bias (a specific "signature" or preferred sequence that the transposase will cut) this is not going to affect the quality of your data. See the paper, it is explained much better than what I did here!
Other interesting details can be found in a review by Reznikoff (Annual Reviews in Genetics, 2008), along with all his previous papers...he´s the master of Tn5 transposase

We have recently made our own version of the protocol (Picelli et al., Genome research 2014) and we found basically the same signature and the same bias as in the original paper. I just prefer the tagmentation approach because is much faster, works great even with pg or sub-pg amounts of DNA and does not require any particular equipment.
Best,
Simone

[Hamid]

Hi,

Since tagmentation does have an inherent insertion bias toward certain preferred sequences, it cannot be considered a random fragmentation system. The other issues with tagmentation are:
1. GC bias. publications have shown that high GC regions are quite under represented in tagmented libraries
2. higher percentage of duplicates in the library as compared to physically sheared PCR-free libraries. The percentage of duplicates tends to increase with the input amount of DNA used.
3. Lower library complexity as compared to PCR-free and physically sheared DNA.
4. Inefficient library preparation when using DNA from FFPE stabilized samples.

Thank you

Hamid

[Simone78]

If input material is not a limiting factor of course the PCR-free approach with Covaris fragmentation is the way to go.
The question is in how many experiments one has micrograms or hundred of nanograms of DNA/RNA available? In my experience the tagmentation, with all its bias, is the only viable option when one has only few picograms of DNA.
Besides, how interested and informative is the expression profile from thousands or million cells?...but this is a topic for another thread maybe...

[luc]

Here is a report of a small scale comparison:

http://pathogenomics.bham.ac.uk/blog/2013/04/adaptor-trim-or-die-experiences-with-nextera-libraries/