Problems with Adapter Dimers

[friedrim]

Hi all,
I have prepared small RNA libraries with the NEBNext kit. In some of my samples, I have a pretty high Adapter dimer contamination (see attached Picture). We have already done AMPure size selection and LabChip (Caliper), but it seems almost impossible to eliminate the dimers since they are very close to our wanted product size (120 vs <130bp).
Do you think that I can still get sufficient read depth with this sample?
Many thanks,
Mona

[nucacidhunter]

Do you think that I can still get sufficient read depth with this sample?

That depends on the number of libraries that are going to be pooled and sequencing platform. You can double the amount of these libraries in the pool and count them as equivalent to two libraries assuming that 50% of reads from those will be from dimers.

[friedrim]

thanks for your answer!

...count them as equivalent to two libraries

Do you mean two of the libraries with dimer contamination Count equivalent to one other library?

We are planning to pool 20 libraries and perform HiSeq single read (50bp).

[nucacidhunter]

Sorry for not posting message clear. I meant that for dimer contaminated libraries you can use twice the amount for pooling and count them as equivalent to two libraries. For instance, after normalizing libraries to 2 nM, one can pool 5 ul of clean library and 10 ul of dimer contaminated ones for use in clustering. So, if 1/2 of the libraries are contaminated you end up with 30 equivalent libraries. Assuming 180M reads output you will obtain 6M clean reads from each small RNA library. That number should be more than sufficient.

[kerplunk412]

Hi Mona,
As nucacidhunter stated, you can simply increase the amount of any contaminated libraries relative to non-contaminated to get enough usable reads, with the caveat that you will have a lot of useless adapter-dimer reads in your final data. For example, if you have 10 libraries that are pure and 10 that are 1/2 adapter-dimer and/or other contaminants, you should use twice as much of the contaminated libraries in your final pool for equal numbers of useable reads.

I am not surprised that Ampure did not work to get rid of the adapter-dimer; I think the inclusion of that method for size selection is pretty optimistic, if not unrealistic, on NEB's part. I have heard some people having good results using the Sage Pippin Prep system, so that is something you may want to look into.

You should be aware that libraries prepared using standard kits, such as that from NEB, suffer from substantial bias that is introduced during the ligation steps. This bias has been shown to be greatly reduced by using adapters with randomized ends. Bioo Scientific recently released a kit using randomized adapters to reduce ligase bias, which can be found here. For more details about this kit or more information on the body of literature exploring causes and reduction of bias in small RNA library prep please feel free to PM me.

For full disclosure, I am an employee of Bioo Scientific.

[NextGenSeq]

Note Ion Torrent miRNA libraries do not show ligation bias, do not have adapter dimer artifacts and do not require gel purification.

I don't work for them.