Multiplexed amplicon library preparation: problems and considerations

[Xavi_]

Hello,

I’m about starting a sequencing experiment with a HiScan SQ.
The point is that we have to prepare the library from multiplexed fragments, varying from 200 to 800 bp, but we are trying to find out how to have a final DNA input where all the amplicons got an equimolecular representation in order to provide a good number of reads and coverage at the end.

Does anyone know how to normalize the effect of mixes with different number of amplicons and concentration? Or this isn’t that important as far as you keep your samples in a good average amplified range?

We tried to get them at the same concentration but it has been impossible. We will have to amplify some of them even in uniplex.

Our aim is to identify SNPs and we are afraid we introduce a bias already before starting the library prep that could decrease the final quality during the analysis.
We usually work with NEBnext and TrueSeq DNA kits and fragment by Covaris method.

I would appreciate very much every help or experience you could provide.

Best regards,
Xavi

[pmiguel]

Ostensibly you can do qPCR titration of each individual library then pool based on your results. We use KAPA (SYBR-green based) kits for this as most do, so we have to adjust our molarities to account for the difference in signal produced by longer amplicons. So you would also need to verify the average size of your amplicons using a gel or Bioanalyzer run. Alternatively you could just design a Taqman assay instead and avoid the need to adjust for amplicon lengths.

As far as pooling amplicons of different lengths in the same lane -- it may not be possible to get even representation of these. From recent results we found when assessing the lengths of inserts of single library pools of a wide size range, it looks like there is heavy bias towards shorter amplicons in any given pool -- even after correcting estimates of molarity for length. Almost like the shorter amplicons reach the flowcell first and longer ones are left trying to find a "seat" among those few remaining. On molecular level, I don't see how this could be the case, though.

My gut feel is that if you had a pool that was 50% 200 bp molecules and 50% 800 bp molecules, >90% of your resulting reads would be from the 200 bp molecules. By far, your best bet would be to pool all the 200 bp molecules in one lane, and the 800 bp molecule in another.

--
Phillip