MiSeq: V1-2 vs V4 for 16S rRNA amplicon sequencing

[fish]

Hi guys,

I'm new to this forum and have a question I couldn't find an answer to.

I plan to sequence fish fin samples to investigate mucosal microbiota. I already run a pilot study with 454 but due to cost issues I want to change to MiSeq.

With 454 I sequenced the V1-V2 region of the 16S and got nice results. For MiSeq most publications present results of V4 sequencing (e.g. Kozich et al 2013). I already asked P. Schloss (mothur) for his opinion and he recommended to use V4 since V1-2 is longer than 250 bp and thus the two reads wouldn't fully overlap what would lead to higher error rates and poor quality data.
In my institute another group develops a protocol for V1-2 for MiSeq to compare their existing results with Roche 454 to the new ones and there it makes sense to use the same region of the !6S, then.

Apart from that I couldn't find a satisfying explanations and would be very happy if you have one.

So, what are the advantages/disadvantages to use V1-2 or V4 of the 16S to sequence on MiSeq?


Thanks in advance!


fish

[mcnelson.phd]

Your question is a good one, and quite honestly there is no good definitive answer because in the end it always comes down to "it depends".

As Pat said, and Rob Knight would agree, the V4 region is good because you can get nearly full overlap of the two reads, which should reduce sequencing noise in your data and thus help prevent OTU inflation. However, the V4 segment is only ~255bp in length, so there's less information contained in a V4 fragment than a longer fragment that spans multiple hyper-variable regions (1-2, 4-5, etc.). I know there's a lot of work being done by different groups on trying to find the best region or combination of regions, but it's very hard to compare the data because it's either all in silico, based on mock communities, or the data was processed differently.

Basically, the V4 protocols as developed by either Rob's group or Pat's group are fairly reliable and both groups have developed and published standard processing methods that should make your analysis fairly easy. The downsides are that because you're working with a small fragment, you may not get as much resolution as you want to tease out OTUs that are different between your samples.

For V1-2, you have a longer fragment that can possibly give you that extra resolution, but at the expense of having to possibly prove that your primers and library construction and analysis methods are valid.

Ultimately, there's a cost consideration in terms of which type of primers do you want to buy in large numbers for library construction. If you have a group on campus that has the primers already made and is willing to share, then that may sway your decision because buying 100 HPLC purified 60-mers is not cheap. As long as you're careful with comparing your results to results generated using another set of amplicons (V4 or otherwise), then you should be fine.

[fish]

Thanks for your reply!

I discussed that issue again with my supervisor and the group leader from the other group and we all agreed that I'll try out the V4. I was informed that the main reason for the others to choose V12 was that there's another project for which that region has to be used and that otherwise they would have chosen V4, either. Further, it might be advantageous to have the primers and stuff for both regions to be able to spontaneously change from one to the other, if necessary.

We also agreed on that there's no general reason to choose this or that region, as you said it depends.

Fortunately, money is not a big issue because I have research money I can invest and additionally we'll use dual indexing and just need 40 primers.

[rhinoceros]

With V12 you could analyze your reads as paired, forward and reverse separately. Expectation would be that all three analyses reveal the same result..

[mcnelson.phd]

Fortunately, money is not a big issue because I have research money I can invest and additionally we'll use dual indexing and just need 40 primers.

Well, if you've got the money and are willing to invest it in primers, then there's no harm in using the V4 like everyone else.

My final thoughts are that with the V3 MiSeq kits now giving you 2x300bp reads, those of us who do 16S surveys really should start looking into longer amplicons because the V4 region is shorter than a single read...

[MadsAlbertsen]

You basically have to test it on your samples. The theoretical evaluation of the "best" primersets are in our experience worth very little.

In all new environments we test V1-3, V3-4 and V4 to see which primerset seem to be best. They are always quite different and its a judgement call to which primers you go for.

With the new MiSeq 2x300 bp kits (and software upgrade) you can easily sequence the entire V1-3 region. The first parts of the reads are excelent quality and the ends overlap so you end with high quality full length V1-3 reads.

Hence, any amplicon < 500 bp should be doable on the MiSeq now.

(Btw, trying 400+300 bp this weekend as it seems doable judgeing from the quality of the MiSeq v3 data we've made so far)

[fish]

Hm, I'll order primers next week and will inform you about my results in couple of weeks.

Thanks to all of you so far!

[mjbuck]

MadsAlbertsen - Can you post your primersets?

[MadsAlbertsen]

MadsAlbertsen - Can you post your primersets?

Sure. I've attached them here. You can find the protocol we use here: http://midasfieldguide.org/

rgds
Mads Albertsen

[firefly2280]

Hi,

I know I'm late to this conversation but we had the same issues. I managed to 'borrow' some Indexed primers for V4 and for V5-7 regions and compare just a few on a small set of my samples. The V5-7 gave better results.

I agree it's hard to know what to do and the best case scenario if you have money, sample and time is to try a few different ones! Sadly, that wasnt possible in my case but it worked out well.

[rhinoceros]

To everyone doing 16S with MiSeq, I very much recommend that you check JGI's related protocols. Your average 454-primers are going to introduce huge bias into 16S MiSeq sequencing, especially so while studying GC-rich communities.

[snetmcom]

To everyone doing 16S with MiSeq, I very much recommend that you check JGI's related protocols. Your average 454-primers are going to introduce huge bias into 16S MiSeq sequencing, especially so while studying GC-rich communities.

Why is that the case? Is it just poor primer design?

[DRYTCYV]

And this one ? - http://supportres.illumina.com/docum...15044223-b.pdf

V3-V4 region

[rhinoceros]

Why is that the case? Is it just poor primer design?

Something about introducing staggered adapters to deal with the low diversity of the ends. They also spike their samples with PhiX. Sorry, I don't recall the specifics. I don't personally deal with the wet part of biology much, so I leave this to those who do

[Richa Sharma]

Illumina Miseq V4-V5 16S rRNAsequencing

Hi all,

This is Richa. I have sent my 16s V4-V5 amplicons for sequencing with illumina MiSeq. But there is a problem. PCR fragments cluster very poorly. There are very less no. of sequences that we are receiving and 50% are phix.

What could be the reason ? How to troubleshoot. Kindly help...!!!!


Thanks.
Richa