Updating metabarcoding protocols for the MiSeq i100

[kbkubow]

Hi Everyone,

I have been doing 16s, ITS, and 18S metabarcoding on an Illumina MiSeq using the one-step Schloss protocol. This relies on combinatorial indices. I recently purchased a MiSeq i100. The MiSeq i100 uses a patterned flow cell, rather then the unpatterned Miseq flow cell, which leads to a much higher incidence of index/barcode hopping. This means that combinatorial indices are really no longer a feasible approach. However, if I stayed with my current one-step PCR approach, this would require purchasing 2x 96 oligos (96 i7s and 96 i5s), which would run me around $2000. And that would only be for one marker (e.g. 16s). I have also been spiking in custom read and index primers, which is more complicated on the new MiSeq i100 and costs more money.

I am trying to decide what the best way is to change my protocol going forward to address both of these issues.

One possibility would be to move to a two-step PCR protocol, where I combine a single primer set that is marker specific for the first PCR followed by a second PCR that adds the barcodes and the Illumina adaptors. Does anyone have protocols for a two-step PCR they recommend? One possibility is to use Illumina's protocol, which uses Illumina unique-dual indices, but there are pretty expensive and are often one-time use only. I would prefer to have my own second step barcoded primers that I could use many times. But it would also be nice if I didn't have to continue using spiked in custom sequencing primers.

My other thought was potentially doing some sort of combinatorial approach, where I order a small number of gene specific unique-dual index primers with tags that would be sequenced within the sequencing reads themselves, say for a half or third of a plate, then pool each half/third of a plate and then make those into Illumina libraries using Illumina's unique-dual index primers. The MiSeq i100 would then demultiplex that data into those pools, and I would then use a downstream program to further demultiplex the samples within each pool using the inline tags/barcodes. Does this make sense? Would this approach still alleviate the issue of index/barcode hopping?

I would love to hear your thoughts on the two possibilities I mention above, or if you have another possibility I haven't thought of. I would also love to be pointed towards any protocols you already know if and would recommend.

Thank you!
Karen

[MU Core]

Hello Karen,

Our group routinely uses a 2 step-PCR approach in preparing targeted amplicon libraries. We have used this flexible approach to allow our researchers to amplify from all sorts of environmental samples. There would be an investment in a 96 primer set. We order the primers mixed from IDT which simplifies the broadcasting into working plates. I've attached our protocol which includes primer design in the appendix.

Best of luck!

Nathan

[kbkubow]

Hi Nathan,

Thank you for this information and the protocol! Just to confirm, with this protocol you do not need to spike in any custom read or index primers? It will work with the standard Illumina sequencing primers contained in the reagent cartridge?

Thank you again!
Karen

[MU Core]

Yes, this protocol works with standard Illumina sequencing primers. We typically use a 15% Phi-X spike-in to overcome low base complexity. A note for this protocol...we control the PCR template input amount closely requiring concentration to be determined by Qubit. We've found that this will normalize the output amount about as well as individual normalization post-PCR. This saves a tremendous amount of time since this allows us to take an equal volume from each reaction and pool for a single bead purification and subsequent QC.

Nathan

[kbkubow]

Sounds good! Noted that normalizing up front saves work downstream. Thank you so much!

[kbkubow]

Hi Nathan,

I was hoping I could ask you for your advice regarding the protocol. I ordered a set of indexed primers and gave the protocol a try, resequencing 24 samples from a previous run using my 1-step PCR protocol. Both were 16s runs. First, I had some issues with having much higher reads in my negative control using the 2-step PCR. Is this me just not being careful enough during the bead cleanups and setting up the second PCR? Also, the majority of my reads were filtered out due to being chimeric when running them through DADA in qiime2, which doesn't happen when I do the 1-step PCR. Do you have any advice for how to minimize the chimeric seqeunces? And also how to minimize the issues with the negative control? I have attached my PCR protocols and the DADA stats for the 1-step versus 2-step PCR runs. I also have two images below, the top one is the 1-step PCR run and the bottom is the 2-step PCR program. Any advice you have would be awesome!

Thank you!
Karen

[HuTasssUn]

Hi Everyone,

I have been doing 16s, ITS, and 18S metabarcoding on an Illumina MiSeq using the one-step Schloss protocol. This relies on combinatorial indices. I recently purchased a MiSeq i100. The MiSeq i100 uses a patterned flow cell, rather then the unpatterned Miseq flow cell, which leads to a much higher incidence of index/barcode hopping. This means that combinatorial indices are really no longer a feasible approach. However, if I stayed with my current one-step PCR approach, this would require purchasing 2x 96 oligos (96 i7s and 96 i5s), which would run me around $2000. And that would only be for one marker (e.g. 16s). I have also been spiking in custom read and index primers, which is more complicated on the new MiSeq i100 and costs more money.

I am trying to decide what the best way is to change my protocol going forward to address both of these issues.

One possibility would be to move to a two-step PCR protocol, where I combine a single primer set that is marker specific for the first PCR followed by a second PCR that adds the barcodes and the Illumina adaptors. Does anyone have protocols for a two-step PCR they recommend? One possibility is to use Illumina's protocol, which uses Illumina unique-dual indices, but there are pretty expensive and are often one-time use only. I would prefer to have my own second step barcoded primers that I could use many times. But it would also be nice if I didn't have to continue using spiked in custom sequencing primers.

My other thought was potentially doing some sort of combinatorial approach, where I order a small number of gene specific unique-dual index primers with tags that would be sequenced within the sequencing reads themselves, say for a half or third of a plate, then pool each half/third of a plate and then make those into Illumina libraries using Illumina's unique-dual index primers. The MiSeq i100 would then demultiplex that data into those pools, and I would then use a downstream program to further demultiplex the samples within each pool using the inline tags/barcodes. Does this make sense? Would this approach still alleviate the issue of index/barcode hopping?

I would love to hear your thoughts on the two possibilities I mention above, or if you have another possibility I haven't thought of. I would also love to be pointed towards any protocols you already know if and would recommend.

Thank you!
Karen

Really interesting transitional challenge and you are definitely not alone in wrestling with the patterned flow cell index hopping problem after upgrading to the i100. The two step PCR approach is probably the most practical path forward for your situation and the Earth Microbiome Project protocol is worth looking at closely since it uses a two step approach specifically designed to be flexible across markers while keeping reagent costs manageable with reusable barcoded primers. Rather than paying Illumina's premium for single use unique dual indices every run, Tropica Casino https://tropical-casino.com/ building your own second step primer set that you own outright and can reuse across multiple markers is a much more sustainable long term solution especially if you are running 16S, ITS and 18S regularly. For your combinatorial inline barcode pooling idea the logic is sound and it would substantially reduce index hopping risk since patterned flow cell hopping affects the sequencing indices but not inline tags read within the actual sequence. Cutadapt combined with a custom sample sheet handles the nested demultiplexing cleanly in most cases. Earth Microbiome Project primers and the Kozich 2013 approach updated for dual indexing are your best starting points before committing to either direction.

[Tiaffan]

Hi Everyone,

I have been doing 16s, ITS, and 18S metabarcoding on an Illumina MiSeq using the one-step Schloss protocol. This relies on combinatorial indices. I recently purchased a MiSeq i100. The MiSeq i100 uses a patterned flow cell, rather then the unpatterned Miseq flow cell, which leads to a much higher incidence of index/barcode hopping. This means that combinatorial indices are really no longer a feasible approach. However, if I stayed with my current one-step PCR approach, this would require purchasing 2x 96 oligos (96 i7s and 96 i5s), which would run me around $2000. And that would only be for one marker (e.g. 16s). I have also been spiking in custom read and index primers, which is more complicated on the new MiSeq i100 and costs more money.

I am trying to decide what the best way is to change my protocol going forward to address both of these issues.
Ridurre l'index hopping richiede una buona preparazione delle librerie e l'utilizzo delle best practice consigliate dal produttore. Per informazioni aggiornate consulto il Naobet Italia sito ufficiale. È sempre meglio verificare i dettagli direttamente dalle fonti ufficiali.
One possibility would be to move to a two-step PCR protocol, where I combine a single primer set that is marker specific for the first PCR followed by a second PCR that adds the barcodes and the Illumina adaptors. Does anyone have protocols for a two-step PCR they recommend? One possibility is to use Illumina's protocol, which uses Illumina unique-dual indices, but there are pretty expensive and are often one-time use only. I would prefer to have my own second step barcoded primers that I could use many times. But it would also be nice if I didn't have to continue using spiked in custom sequencing primers.

My other thought was potentially doing some sort of combinatorial approach, where I order a small number of gene specific unique-dual index primers with tags that would be sequenced within the sequencing reads themselves, say for a half or third of a plate, then pool each half/third of a plate and then make those into Illumina libraries using Illumina's unique-dual index primers. The MiSeq i100 would then demultiplex that data into those pools, and I would then use a downstream program to further demultiplex the samples within each pool using the inline tags/barcodes. Does this make sense? Would this approach still alleviate the issue of index/barcode hopping?

I would love to hear your thoughts on the two possibilities I mention above, or if you have another possibility I haven't thought of. I would also love to be pointed towards any protocols you already know if and would recommend.

Thank you!
Karen

A two step PCR approach sounds like the most practical long term solution. It separates target amplification from indexing, reduces primer costs across multiple markers, and works well with unique dual indexing on patterned flow cells. Inline tags can work, but they add extra demultiplexing complexity and still require careful validation to ensure index hopping does not create sample misassignment. If possible, I would test both methods on a small pilot run before fully switching protocols.

[kbkubow]

Hi Everyone, I appreciate all of your thoughts and suggestions! I ended up using the 2-step protocol shared by Nathan, and it has been working well. I have used it for both 16S and ITS. At first I was confused about why I was getting so many chimeric sequences in the downstream analysis, but then I realized it had to do with trimming the primer sequences. In the 1-step Schloss protocol we spiked in custom sequencing primers which ensured that the sequencing reads started after the primer sequences. Using the 2-step protocol with the standard Illumina sequencing primers means that Read 1 and Read 2 sequencer the primers as well as the amplified sequences. I didn't realize this at first. Once I trimmed the reads to remove the primers, I had many fewer chimeric sequences and everything worked well. Just a heads up for anyone else making the switch!