Pooling and sequencing NexteraXT and TrueSeq PCR free libraries on a single lane

[sinnafoch]

Hi everyone,

I have a question regarding the pooling of NexteraXT and TrueSeq PCR free libraries onto a single Illumina HighSeq lane. I am sharing a lane with a colleague and we have different needs: I have enough DNA so I would like to use PCR free libraries. My colleague on the other hand has very little DNA and wants to use NexteraXT preps. Each of use would like to use 1/2 of a lane for sequencing.

The person at our sequencing facility told us that pooling PCRed and PCR-free libraries could be problematic because it is hard to balance molarities in such a way that the different library types each occupy 1/2 of a lane.

Now we are wondering about the possible reasons behind this? Does anyone have an idea why this could happen. I have never seen this mentioned in a paper.

To which extent is this really a problem? When aiming for a 50:50 lane split how far off could the output be? 60:40? 80:20?

thank you all for your help!

[GenoMax]

What are the expected insert sizes in both cases? Keep in mind that in general short insert libraries will preferentially cluster (you may need to adjust loading conc going in, taking that into account) and may compete out larger fragments.

[Qibao]

It also depends on what instrument you are using. We tried it on HiSeq4000 before and it worked well. It uses 2-3nM for PCR based libraries and 1-2nM for PCR free libraries. We mixed the samples to 3:2 ratio and it gave us expected sequencing percentage.

[Genetic Librarian]

well, mixing two libraries is a tricky task in general, no matter if PCR-free or not.
It works well with libraries of the same fragment size. If the two libraries are of different sizes, there is a certain amount of educated guessing to counteract the bias for smaller fragments.

For your particular issue, you need to make sure that both libraries are quantified using the same methods. For PCR free libraries, qPCR is needed, whereas for Nextera XT you are fine with just Qubit. So in this case, both should be measured with qPCR.

[sinnafoch]

Thanks everyone! That the libraries probably need to be quantified with the same method is a good hint, we have also thought that this may minimize problems in our case.
We have not thought about insert size however, I don't really understand why this may be an issue.

Anyways, thank you again for your replies, gives me something to think about

[Genetic Librarian]

Insert size is a problem because bias works in favour of smaller fragments.
So if you mix a 300 bp library and a 500 bp library 50/50, you will end up with a ~70/30 read distribution.
One reason is the normal PCR bias during bridge amplification.
Another reason is the ExAmp clustering reaction specific to the HiSeq 3000/4000. In the ExAmp reaction, the reaction mix is extremely viscous. Apparently, this should limit DNA motility so that the chance that multiple fragments attach inside the same nanowell is reduced. Basically, the ExAmp reaction relies on the fact that library binding is much slower than library amplification. So, you could think of the reaction mix like of an agarose gel, where smaller fragments move faster. Thus, smaller libraries will populate and block a proportional higher number of nanowells on the patterned flowcell.

[sinnafoch]

Ok thank you for the explanation. This makes it more clear.

[finswimmer]

We have not thought about insert size however, I don't really understand why this may be an issue.

Beside the mentioned reason: If you have two librarys with the same concentration but with different fragment size, this also means, that the total number of fragments in the library with the lower fragment size is higher. And every fragment is apotential read.

fin swimmer

[Genetic Librarian]

Beside the mentioned reason: If you have two librarys with the same concentration but with different fragment size, this also means, that the total number of fragments in the library with the lower fragment size is higher. And every fragment is apotential read.

That should not be the case, because you pool equimolar.

[finswimmer]

That should not be the case, because you pool equimolar.

You're right. I didn't have a coffee at the time of writing

fin swimmer