Fusion Primer Vs. Ligation of Adaptors

[cmuletz]

Hi all,

I was recently informed from a 454 Core Facility that "amplicon libraries do not sequence as well as libraries where the adaptors are incorporated via ligation (50-100K reads vs. 150-200K reads per ¼ run, for example)." They were referring to using the new Y-adaptors.

I found this surprising. From only thinking about this, it seems that doing the PCR reaction with the adaptors and everything there would be better than ligating the adaptors on. My reason for this is that you only need to do 1 PCR reaction for the fusion primer while for the ligation you would have to do 2 (one to amplify products and one after ligation to amplify products with adaptors), so more potential for PCR error. I tried to do a literature search, but found nothing.

Any help would be great.

[RCJK]

I think they may be referring to shotgun libraries (rapid libraries) where the adaptors are ligated on and not necessarily amplicons with ligated adaptors. Amplicon sequencing typically does have a lower yield compared to rapid libraries.

[ajthomas]

I'm surprised at their statement. In my experience, amplicon libraries produced with fusion primers generally sequence better and there is less variability between libraries. When libraries are prepared that way, every molecule is functional in emPCR, whereas libraries that are prepared by ligation can vary because of the efficiency of the ligation reaction. There isn't any second PCR after the ligation, but if you were to do that, you would end up the same as if you had used fusion primers.

The reason amplicons generally yield lower numbers of reads than shotgun libraries is because of the processing afterward. Even then, there should not be as much difference in as they are telling you. I usually see maybe 10-20% fewer reads when processing with the amplicon pipeline compared to the shotgun pipeline. If you process amplicon libraries with the shotgun pipeline you will get just as many reads as you do with a shotgun library. The amplicon pipeline throws out more reads because it has some extra filters that are applied that help with amplicon data but would not be good to use with shotgun data. The actual sequencing works just as well.

If they are telling you that you will get more reads if the adapters are ligated on, then they are not processing the data using the amplicon pipeline. Amplicon libraries should be processed using the amplicon pipeline whether they are prepared with fusion primers or by ligating the adapters on. You could ask them to process the data both ways and give you both sets of data so you can determine for yourself which is better for you.

[NathalieA]

I found this discussion very interesting.
In my compagny, we usually used fusion primers for amplicon libraries....and we have no problems. as it said before, the number of reads depends on the prossesing afterwards. We used only the shotgun process even if we have amplicon libraries.

but since one month we have a big project with amplicon library we need to use ligation (the customer ask us to do like this! 2 step of PCR, and after that ligation). And now we have some problems... libraries are really heterogeous in emulsion... we saw that RL quantification by Roche is poor quality.
And moreover, we saw that several MID have a very few sequences regards the others... we confirm this observation through 2 pools and 2 runs...

do you observe this MID effet with the new protocol of RL?

[cmuletz]

You may want to contact Beckman Coulter Genomics to see if they will be helpful. I did not end up ligating the 454 adaptor sequences, but chose to use fusion primers.