Preparation for amplicon sequencing without PCR

[GMDickson]

Hi all,
I have had a good look through the forum posts and don't seem to find the same questions I have about preparing amplicons for sequencing (on a MiSeq) without using a 2-step PCR.

I need to sequence short amplicons from degraded DNA in a metabarcoding experiment. I have been advised repeatedly to avoid the 2-step PCR due to artefact and downstream issues. Therefore, I was wondering if the following approach would work or if there are any major do's and don't in relation to this:
1. amplifying using indexed (and variable N incorporated - it will be very low complexity) target sequence primers (in the presence of a blocking oligo to remove non-target amplification)
2. ligate P1/P2 and sequencing adaptors together (??)
3. Use biotinylation to remove unwanted P1/P1 fragments.
I see that without the PCR there is no enrichment step in this process but other than that (and therefore the clean-up), is there any way to avoid PCR steps in preparing amplicons for sequencing?

I have read the 16S and other metabarcoding protocols but they all seem to add adaptors to amplicons via PCR. I know that one commercial lab uses a ligation process but can't get details of it and I need to complete this in house, ideally without using any expensive kits. Does anyone know of a protocol that would be useful in this situation or if this query has been addressed elsewhere (in which case I would be most, most grateful for a link to it).

Many thanks in adavnce for any advice on this.

[HESmith]

I was wondering how you'd generate amplicons w/o PCR, but see that one round is acceptable. You can incorporate the sequencing adaptor ends into your target sequence primers, with random k-mers between the two segments to discriminate PCR duplicates.

[GMDickson]

Prep of samples for MiSeq without 2nd PCR query

Hi HESmith,

Many thanks for your response.

Yes, we are using PCR to produce the amplicons initially but would prefer to avoid further PCRs usually necessary to attach the sequncing adaptors/cleanup/enrich. I have indexes and a variable number of Ns added to the end of my PCR primers so that I can distinguish individual samples and try avoid problems associated with low complexity samples but I am not sure if it is feasible to add the necessary sequencing (for MiSeq) adaptors on by ligation or whether a portion needs to also be included in the initial PCR primers (similar to a protocol I found for preping samples for the Ion Torrent). I need to multiplex samples in a run and am dealing with very low quality DNA so the fragment is small.

Do you know of any protocols that might describe the addition of sequncing adaptors without a PCR step?

Many thanks again for your help.

[nucacidhunter]

Yes, we are using PCR to produce the amplicons initially but would prefer to avoid further PCRs usually necessary to attach the sequncing adaptors/cleanup/enrich. I have indexes and a variable number of Ns added to the end of my PCR primers so that I can distinguish individual samples and try avoid problems associated with low complexity samples but I am not sure if it is feasible to add the necessary sequencing (for MiSeq) adaptors on by ligation or whether a portion needs to also be included in the initial PCR primers (similar to a protocol I found for preping samples for the Ion Torrent). I need to multiplex samples in a run and am dealing with very low quality DNA so the fragment is small.

Do you know of any protocols that might describe the addition of sequncing adaptors without a PCR step?

You can use any DNA library prep kit with forked adapters to prepare PCR-free library from amplicons. You would need to avoid size selection step and just do couple of clean ups after ligation to remove excess adapters.

Generally using non-template Ns and barcode at 5' end of PCR primers will result in bias resulting in quantitative and and sometimes qualitative differences deppending on the barcode.

[nucacidhunter]

You can use any forked-adapter ligation based DNA library prep kit to prepare PCR-free libraries with some modification in protocols.

Generally adding Ns and barcodes to 5' end of PCR primes results in bias that can be quantitative and sometimes qualitative making it less reliable for 16S studies.

[GMDickson]

Prep of samples for MiSeq without 2nd PCR query

You can use any forked-adapter ligation based DNA library prep kit to prepare PCR-free libraries with some modification in protocols.

Generally adding Ns and barcodes to 5' end of PCR primes results in bias that can be quantitative and sometimes qualitative making it less reliable for 16S studies.

Hi Nucacidhunter,

Thanks for the info. I've been looking at the forked adaptor to try and increase the proportion of P1/P2 fragments (or P5/P7) but I was wondering if it is possible to just ligate these without using a kit.

I think that the TruSeq kit for PCR free prep looks closest to what I need except that we already have combinations of our own indexes to use. We have a lot of samples, need to multiplex and process with very limited funds so I don't think we can use a kit.

I was hoping that I could add the indexes and Ns to the primers, then ligate an adaptor and sequencing primer - I am just not sure whether there are published sequences that I could just ligate before cleanup and sending to the sequencer, or whether I definitely need to include an adaptor before annealing the sequencing primers.

Many thanks again for taknig the time to comment and any further thoughts or advice are much appreciated.

[HESmith]

Ligation is always less efficient than incorporating the adapter ends directly into your amplicon primers. Is there a reason NOT to use the following type of primer?

adapter end - index - random k-mer - target sequence

If both primers have this structure, you can use dual indexing, alter the k-mer length to address phasing, and distinguish PCR duplicates, with only one round of PCR.

[GMDickson]

Ligation is always less efficient than incorporating the adapter ends directly into your amplicon primers. Is there a reason NOT to use the following type of primer?

adapter end - index - random k-mer - target sequence

If both primers have this structure, you can use dual indexing, alter the k-mer length to address phasing, and distinguish PCR duplicates, with only one round of PCR.

Thanks HESmith. This was the original idea I have been toying with.

The thought of ligating the adaptors on latterly was really just to keep costs down for the many samples we will need to run. I have heard that usual way this is done for metabarcoding is to PCR using a primer that includes the random k-mer + index + target sequence primer, and then ligate on the sequencing adaptors (avoiding further PCR). However, my understanding is that the latter part of this protocol is usually done by a commercial company and we will not have the resources to use their services in the future. Therefore I've been trying to develop a protocol that will be as cost effective (yet simple!) as possible to do inhouse.

I'm really grateful for your comments as they are allowing me to start ticking or scratching out some of my ideas I've had on how to tackle this. I have been going around in circles!

Thanks again.

[HESmith]

Yes, the longer oligos are more expensive. But with dual indexing, 20 primers can accommodate 100 samples.

One more point to consider: PCR-free ligation prep requires a substantial amount (1-2 micrograms) of input DNA so, to generate that amount of amplicon in one round of PCR, you may be introducing the bias that you're trying to avoid.

[GMDickson]

One more point to consider: PCR-free ligation prep requires a substantial amount (1-2 micrograms) of input DNA so, to generate that amount of amplicon in one round of PCR, you may be introducing the bias that you're trying to avoid.

I know I am worried about that as well but not sure how to avoid it. All of the samples will be of the similar but I am trying to work out how much product we are getting out from the first PCR at the moment to see how much of an issue this is.

Many, many thanks for your help. I am sure I will be back with more ponderings (and any results I have) shortly.

Cheers.

[nucacidhunter]

1-2ug input is recommended for shotgun libraries that are sequenced deep and should have high diversity. Amplicon libraries can be sequence deep but they are low diversity so input can be reduced to 100ng and also library prep can be scaled down to reduce costs.

Generally amplicon sequencing with Illumina systems can utilise following methods:
1- Fusion primers which include target specific sequences+ linker+ padding+ index+ flow cell binding domain such as http://press.igsb.anl.gov/earthmicro...standards/16s/
2- Two step PCR as demonstrated in Illumina 16S protocol
3- Ligation of amplicons (which can include sample specific barcode at 5' end but potentially will affect quantitative results) with indexed adapters

For low diversity libraries various methods can be adapted to increase base diversity during sequencing. For qualitative application their effect would be minimal but for quantitative assays such as 16S microbial profiling diversity addition methods will affect the results.

[GMDickson]

Many, many thanks nucacidhunter. This is reassuring. I think we will try for approach 3 as we were not expecting qualitative results. They would be great but for now I don't think they will be possible.

Many thanks again for your help.

[nucacidhunter]

Happy to help and this marks my 1000 post.

[GMDickson]

Happy to help and this marks my 1000 post.

You should get a badge of thanks!