Re-enrichment?

[aplany]

Hi all. I'm constructing libraries for Illumina 100bp paired-end sequencing. I work on a non-model organism, and a large amount of good RNA is difficult to get. Library construction following the TruSeq protocol with Covaris shearing and a size selection step with a gel prior to enrichment has resulted in libraries with very low concentrations - between 4.5ng/uL and 8ng/uL. This is most likely because of the low amount of starting material for the library prep (between 500ng and 3ug). My question is - can I re-enrich this end-product prior to sequencing, or will it create problems?

Any other suggestions would be appreciated.

Thanks!

[Wurstmensch]

Hi aplany. I had benefit a lot from this forum, now I try to give something back.

In my opinion it isn't a good idea to re-enriche your library. Too many cycles of PCR can result in a bios of your library. As a result of this you have the risk of amplifying mistakes of the DNA-Polymerase. With more PCR you have the possibility to create a higher concentrated library but get results out of the sequencing that you can't trust.

I'm relatively new to this topic, so please correct me if I'm wrong, but I had the same discussion in my workgroup about my own library and we decided, that more than 15 cycles of PCR are to riskfull.

[pmiguel]

Why? 4.5 ng/ul would put you somewhere north of 15 nM, right? (Possible factor of 2 error in that calculation). That should be plenty. Unless I am missing something.

What was your quantitation method?

--
Phillip

[niceday]

I'm with pmiguel. Your library sound good. Even if you are slightly below 10 nM you can alter your protocol slightly and get good clusters. Too many cycles gives too much redundancy.

[aplany]

Hi all - thanks for the help. The quantitation method was nanodrop. I quantified again using QBit and this time the concentrations are between 1.45 and 3.44ng/uL. Seems like some of the libraries are usable - can get 10nM. A few are below 10, though.

Can I pool 5nM of one library with 10nM of another for multiplexing in a single lane or would that mess things up? Would the whole sequencing run have to be run with amounts of 5nM?

[seq_me]

Aplany, I am trying to answer your question in three parts. third part is main explaination.

You can mix pool libraries at different concentration. Theoreticall you will get the reads in the same ratio as well.

You can run the libraries at different concentration.

I would suggest to look in the cBOT user guide to look in the template prep to understand qhat minimum quantity of the libraries you can get away with. This follows from there. If you are loading the library at a final concentration of 10 pM you need 120 ul of it.(So theoretically for a 2nM lib you need 0.6 ul) In general as per the protocol you are using 120 ul out of the 2000 ul that you can make from the 2nM/10 ul of the library. Hence there is a good margin to load libraries at concentrations lower than yours.

[pmiguel]

Is this a GA-IIx? If so, no idea.
For a HiSeq/HiScanSQ using v3 chemistry you should be able to get away with a concentration as low as 1.2 nM if you are shooting for a final concentration of 20 pM.

There is always the issue of what % of your molecules are actually amplicons. But unless you do qPCR, that is not easy to say until you see your clustering results.

So, if your enrichment PCR results produced less library than you expected, then there may have been an issue during library construction that hosed the library. So you might want to at least do a qPCR to verify your concentration.

--
Phillip

[seq_me]

My reply is based on using GaIIx