SEQanswers

Go Back   SEQanswers > Applications Forums > Sample Prep / Library Generation



Similar Threads
Thread Thread Starter Forum Replies Last Post
1000bp insert size with Illumina TruSeq DNA PCR-Free Library prep kit Lovro Sample Prep / Library Generation 13 10-03-2016 11:10 AM
PCR step in DNA library prep? liumangmang Illumina/Solexa 5 10-09-2014 02:16 PM
Amplicon PCR Problems with Illumina MiSeq Sample Prep locon833 Sample Prep / Library Generation 4 08-28-2014 12:52 PM
Multiplicom MASTR Dx - 2nd round of PCR with different polymerase jana.supikova Illumina/Solexa 0 01-29-2014 06:19 AM
Fluidigm PCR amplicon Library prep yog77 Sample Prep / Library Generation 2 10-30-2012 04:36 AM

Reply
 
Thread Tools
Old 11-30-2016, 05:23 PM   #1
shreve697
Junior Member
 
Location: Pennsylvania

Join Date: Feb 2016
Posts: 6
Default Help with first round PCR optimization for Illumina MiSeq Metabarcoding Library Prep

I am currently optimizing PCR for amplification of the ITS2 region for fungi and V4 region for archaea/bacteria from environmental gDNA. This amplification is being done using degenerate primers appended with Illumina Nextera adapters (as detailed below). The PCR product from this first round will be submitted to a sequencing facility on campus for a second round of PCR where indices will be added, then sequencing will be carried out on an Illumina MiSeq.
I have encountered two issues:

1) After running an annealing temperature gradient, I am still uncertain which annealing temperature to use for the V4 primer set.

2)
Both ITS2 and V4 primer sets are yielding non-specific product which show up as a faint smear with some diffuse 'bands' in my gel.
I've attached a gel image of the Ta gradient (annotated in PowerPoint) which depicts both issues.

For issue (1),
my Ta gradient yielded virtually identical bands across more than 20C span of annealing temperatures – from 50C all the way up to 72C. This makes it difficult to select an annealing temperature based on the gradient results. Based on these results, I could select a Ta as high as 72C and run a two-step protocol, but I’m worried that I may miss some taxa by selecting a Ta which is so far above the lower Tm. I am using a popular V4 primer set (515f-Y/806rB, see below) that has a large difference in melt temperatures. Under the buffer conditions used, the Tm of the forward primer is 74C, whereas the Tm of the reverse primer is only 56.6C.

Does anyone have experience with this (degenerate) primer set?

Is it expected to get similar amplification over such a wide range of annealing temperatures?

Why do so many people use a primer set that does not conform to the basic rules of primer design, yet nobody talks about this in the literature or elsewhere?

For issue (2),
both the ITS and V4 product has non-specific smears at higher molecular weights above the expected bands. For ITS some of this may be due to varying ITS length in the potentially complex community from which the gDNA was derived, but the V4 product is expected to be ‘cleaner.’ I tried titrating the primer and shortening the extension time for the ITS reaction to little affect, but haven’t had time to troubleshoot the V4 reaction yet. I suppose I would start by titrating the primer concentration. The cycle count is high, but I would like to optimize at the highest expected cycle count assuming that lower cycle counts will yield cleaner product.

I’m wondering if anyone has experience with these primer sets (especially degenerate versions used here) and whether this amount of non-specific product is common?

Is this caused by degeneracy in the primers?

I anticipate that the presence of non-specific product will reduce the number of usable reads in my sequencing results at the very least. Should I be concerned about this, or proceed with the second round of PCR and sequencing?

PCR Protocol:

Polymerase: Invitrogen Platinum SuperFi 2X MM

Recipe:
MasterMix (2X) 5 uL
Forward primer (10 uM) 0.5 uL
Reverse primer (10 uM) 0.5 uL
Tmeplate (10 ng/uL) 1 uL
PCR Water 3 uL
Total Reaction volume 10 uL

Thermocycler Program (based on Invitrogen suggested protocol for SuperFi polymerase):
Cyc1 (1X) 98C for 2 min
Cyc2 (35X) 98C for 10 s
(50-72C) for 10s
72C for 15 s
Cyc3 (1X) 72C for 5 min
Cyc4 4C HOLD

Primers:
All primers have the following Illumina adaptors appended:
Forward – TCGTCGGCAGCGTCAGATGTGTATAAGAGACAG
Reverse – GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAG

Tm listed below for each primer was calculated at www.thermofisher.com/tmcalculator

For ITS,

ITS3mix (forward)
CATCGATGAAGAACGCAG 61.1
CAACGATGAAGAACGCAG 61.2
CACCGATGAAGAACGCAG 63.7
CATCGATGAAGAACGTAG 54.5
CATCGATGAAGAACGTGG 60.8
CATCGATGAAGAACGCTG 61.1

ITS4ngs (reverse)
TCCTSCGCTTATTGATATGC 62.5

Tedersoo, L., Anslan, S., Bahram, M., Põlme, S., Riit, T., Liiv, I., … Abarenkov, K. (2015). Shotgun metagenomes and multiple primer pair-barcode combinations of amplicons reveal biases in metabarcoding analyses of fungi. MycoKeys, 10, 1–43. http://doi.org/10.3897/mycokeys.10.4852

For V4,
515f-Y (forward)
GTGYCAGCMGCCGCGGTAA 74

806rB (reverse)
GGACTACNVGGGTWTCTAAT 56.6

Walters, W., Hyde, E. R., Berg-lyons, D., Ackermann, G., Humphrey, G., Parada, A., … Jansson, J. K. (2015). Transcribed Spacer Marker Gene Primers for Microbial Community Surveys. mSystems, 1(1). http://doi.org/10.1128/mSystems.00009-15.Editor
Attached Files
File Type: pdf Annotated gel images - Platinum SuperFi Gradient2.pdf (238.5 KB, 14 views)
shreve697 is offline   Reply With Quote
Old 12-01-2016, 12:06 AM   #2
nucacidhunter
Senior Member
 
Location: Iran

Join Date: Jan 2013
Posts: 991
Default

General comments:

1- The lowest annealing temperature in the gradient that does not produce non-specific amplicons should be chosen to minimise potential biases.
2- If cycle number is high the yield with lowest and highest annealing temperature could be similar but once you prep, sequence and analyse data the differences would be identified.
3- The limitation of primers is known. The important factor is reproducibility and not perfection.
4- If you can visualise amplicons in gel, probably they have been over amplified and hence you will see high molecular amplicons resulting from fusion.
5- 10 ng input for a 10 ul reaction is too high and probably some of HMW seen on the gel is input DNA.
6- Large fragments are not a major issue because smaller amplicons will be sequenced preferentially.
7- Should increase annealing time to 1 min for lesser bias because they are degenerate primers.
8- Non-specific products will be discarded during analysis so they do not have large impact.
9- PCR cycles should be kept below 30
nucacidhunter is offline   Reply With Quote
Old 12-01-2016, 11:40 AM   #3
shreve697
Junior Member
 
Location: Pennsylvania

Join Date: Feb 2016
Posts: 6
Default

Thanks for the quick response, nucacidhunter.

A couple of follow up questions:

1) So, I should not be able to see the amplicon on the gel? Or do you mean to suggest that the expected band should only be very faint?

2) By how much should I reduce my target input (as a starting point)?

3) Is there any concern about using such long annealing time with a highly processive polymerase? i.e. Since the enzyme is highly processive, might we see significant activity over the longer time despite the lower temperature? Just a hesitation, since all recommended cycle times are significantly shorter for this polymerase than for taq.

Thanks for taking the time..
shreve697 is offline   Reply With Quote
Old 12-01-2016, 10:45 PM   #4
nucacidhunter
Senior Member
 
Location: Iran

Join Date: Jan 2013
Posts: 991
Default

1- I don't know how much of 10 ul reaction has been loaded but generally you need just enough to know that PCR has worked.

2- Input quantity depends on the portion of microbial DNA in the sample and samples collected from human saliva or animals will have high ratio of host DNA which will require more input. I would not go above 0.3 ng/ul of reaction.

3- I have not looked at that polymerase specification but you are better off with longer annealing time to reduce biase. In first instance best approach would be a trial and looking at amplicon profile and yield.

From first PCR you only need enough yield to visualize amplicons and 4 ng for input into second PCR which would yield over 100 ng library by 5 cycles.
nucacidhunter is offline   Reply With Quote
Old 12-05-2016, 05:48 AM   #5
shreve697
Junior Member
 
Location: Pennsylvania

Join Date: Feb 2016
Posts: 6
Default

Thanks for sharing your wisdom, nucacidhunter. Much appreciated!
shreve697 is offline   Reply With Quote
Old 12-05-2016, 06:40 AM   #6
thermophile
Senior Member
 
Location: CT

Join Date: Apr 2015
Posts: 192
Default

Generally nucacidhunter is correct that you need to increase your annealing time for degenerate primers and complex template. But some of the new error proofing polymerases are very aggressive and will "error correct" everything leading to non-specific amplification if given too long of annealing time. I don't know if the one that you are using falls into that category, check with tech support.
__________________
Microbial ecologist, running a sequencing core. I have lots of strong opinions on how to survey communities, pretty sure some are even correct.

Last edited by thermophile; 12-05-2016 at 06:46 AM.
thermophile is offline   Reply With Quote
Old 12-07-2016, 04:36 PM   #7
shreve697
Junior Member
 
Location: Pennsylvania

Join Date: Feb 2016
Posts: 6
Default

Thanks for the heads up, Thermophile. I'll contact Tech Services to inquire on the maximum recommended annealing time.

In the meantime, I reduced the template input (from 10 ng/10 uL rxn to 2.5 ng/10 uL rxn) and number of cycles (from 35 to 25) as Nucacidhunter had suggested.

My results were definitely improved (as summarized in the attachment along with gel images); however, another issue has arisen. On the initial gradient where I had HMW smears, Ta selection was difficult for my bacterial V4 primer set due to the products looking about the same for each annealing temperature. However, the fungal ITS primers showed the best amplification at an annealing temperature very close to that calculated by a polymerase/buffer specific Tm calculator at www.thermofisher.com/tmcalculator.

After reducing the number of cycles and template input, the smearing is largely gone, except at the lowest annealing temperatures for the V4 primers. In fact, the best looking amplification products for V4 now coincide with the calculated Tm of the lowest Tm primer. Unfortunately, the results for the ITS primer set are now a bit confusing. Now I am only getting bands at the lowest annealing temperatures, and the best looking product is at an annealing temperature a full 8 degrees below the calculated one.

I know that the calculated Ta is a rough estimate at best, but the fact that it is so far off is concerning. I can also rationalize the losing visible bands at higher annealing temperatures as I reduce the number of cycles.

I guess my question is: If the gradient PCR reveals a different optimal Ta based on number of cycles and template input, how do I know what the next step should be? For instance, I could increase the number of cycles or template input for the ITS amplification in order to get product at a temperature that is closer to the expected one, but that seems a bit forced.

Any comments on my gel images or suggestions about where to go from here would be much appreciated.

Thanks for your time, all.
shreve697 is offline   Reply With Quote
Old 12-07-2016, 08:21 PM   #8
nucacidhunter
Senior Member
 
Location: Iran

Join Date: Jan 2013
Posts: 991
Default

The next step could be trying lowest and the optimum annealing temperature based on gradient PCR with standard DNA (Zymo Research offers ZymoBIOMICS microbial community standard DNA with known composition) and some real life samples with at least 3 replicate for each and also 25 and 30 PCR cycles. For instance, 3 template (1 standard+ 2 real samples) with 2 annealing temperature and 2 different cycling (25 and 30) for each with 3 replicates (3x2x2x3= 36 combinations). These amplicons can be used for final library prep, one MiSeq sequencing and analysis.

Most important factor in 16S or ITS based profiling is reproducibility. I would choose the lowest annealing temperature and the highest cycle or in between those two (to ensure success when target template in sample is low) that accurately identifies know sample species and gives reproducible results for all samples.
nucacidhunter is offline   Reply With Quote
Old 12-08-2016, 05:51 AM   #9
thermophile
Senior Member
 
Location: CT

Join Date: Apr 2015
Posts: 192
Default

Do you have access to another thermocycler? changing input amount shouldn't cause drastic annealing temperature changes. Are you doing these 10ul rxns in 384well plates or 96?

In optimizing, I'd use 30 cycles. somewhere around 37-38 you are going to be reliably amplifying your kitome.
__________________
Microbial ecologist, running a sequencing core. I have lots of strong opinions on how to survey communities, pretty sure some are even correct.
thermophile is offline   Reply With Quote
Old 12-09-2016, 01:36 PM   #10
shreve697
Junior Member
 
Location: Pennsylvania

Join Date: Feb 2016
Posts: 6
Default

I've attached a Powerpoint file with all of my gradients thus far laid out visually in a grid (cycle count x template input). The results show that the template input had a minor impact on which annealing temperatures had a visible band, but that effect leveled off at higher template inputs. You can see I tested some pretty high template concentrations for ITS, but that is because I suspect that fungi are a minor part of the community and most gDNA is from bacteria. For both V4 and ITS, cycle count appears to have a much bigger effect on which annealing temperatures have visible bands. Is it expected that cycle count would have such an effect? It kind of makes sense to me, but I've not seen this mentioned anywhere.

General thoughts on my gradients and selection of reaction conditions would be much appreciated.

Nucacidhunter, the scheme you've laid out sounds good, but I'm constrained by time and money and unable to make a selection based on a sequencing run. I will, however, be running a few samples through the second round of PCR (to attach Nextera indices) and then through a bioanalyzer to check amplicon size distribution prior to processing the rest of the samples and sequencing.

You mention reproducibility - I have seen where some authors have run each sample at different template input and cycle counts in order to achieve a "medium intensity" band on the gel for each sample. Is that a reasonable approach in your opinion, or would it be best to run all samples at the same template input and cycle count?

Thermophile, I have only the one thermocycler which will run a gradient. There are certainly others on campus, but they will likely be quite a ways from my lab. I'm doing the 10 uL reactions in 200 uL PCR tubes (equivalent to 96 well plate, I suppose). The thermocycler I'm using (Gstorm) allows me to input the reaction volume, so I assume it is adjusting the ramp rate, etc., accordingly.
Attached Files
File Type: ppt Grid Summary of ITS Annealing Gradients.ppt (1.66 MB, 8 views)
shreve697 is offline   Reply With Quote
Old 12-12-2016, 09:02 AM   #11
thermophile
Senior Member
 
Location: CT

Join Date: Apr 2015
Posts: 192
Default

I use similar primers

515 GTGCCAGCMGCCGCGGTAA
806 GGACTACHVGGGTWTCTAAT

and anneal at 55, when I've done gradient PCR for these primers I've not seen any amplification above ~62. I do 15ul rxn in 384 well plates with ~10ng template, so not as high as you but close. I still think you have an uneven temperature issue, try doubling or even tripling your rxn volume and see if that changes this. You could also try out "thin walled" tubes if you can find them.
__________________
Microbial ecologist, running a sequencing core. I have lots of strong opinions on how to survey communities, pretty sure some are even correct.
thermophile is offline   Reply With Quote
Old 12-13-2016, 02:16 AM   #12
nucacidhunter
Senior Member
 
Location: Iran

Join Date: Jan 2013
Posts: 991
Default

Quote:
Originally Posted by shreve697 View Post
I've attached a Powerpoint file with all of my gradients thus far laid out visually in a grid (cycle count x template input). The results show that the template input had a minor impact on which annealing temperatures had a visible band, but that effect leveled off at higher template inputs. You can see I tested some pretty high template concentrations for ITS, but that is because I suspect that fungi are a minor part of the community and most gDNA is from bacteria. For both V4 and ITS, cycle count appears to have a much bigger effect on which annealing temperatures have visible bands. Is it expected that cycle count would have such an effect? It kind of makes sense to me, but I've not seen this mentioned anywhere.

General thoughts on my gradients and selection of reaction conditions would be much appreciated.

Nucacidhunter, the scheme you've laid out sounds good, but I'm constrained by time and money and unable to make a selection based on a sequencing run. I will, however, be running a few samples through the second round of PCR (to attach Nextera indices) and then through a bioanalyzer to check amplicon size distribution prior to processing the rest of the samples and sequencing.

You mention reproducibility - I have seen where some authors have run each sample at different template input and cycle counts in order to achieve a "medium intensity" band on the gel for each sample. Is that a reasonable approach in your opinion, or would it be best to run all samples at the same template input and cycle count?
Looking at the grid I think 0.25ng/ul input with 28 cycles and annealing at 50C would give good results. To optimise annealing and extension time I would first try increasing extension to 30 sec and if the results are similar to 15 sec then try increasing annealing time to 20 sec and compare results. One should use maximum extension time that does not affect the results negatively because shorter extension time can result in partially extended templates and increase the non-specific and fusion amplicons ratio. Similarly, increasing annealing time and lowering temperature ensures that priming is unbiased and specific.

I think the same condition should be used for all samples that will be compared unless someone can show with a known standard community DNA that using different conditions give similar results.
nucacidhunter is offline   Reply With Quote
Old 12-13-2016, 05:57 AM   #13
thermophile
Senior Member
 
Location: CT

Join Date: Apr 2015
Posts: 192
Default

I think it's ok (preferable) to optimize ITS and 16s separately. It's very difficult to quantitatively combine bacterial and fungal communities for analyses.

accuprime II we use 30sec annealing, 1min extension, 10min final extension for both v4 16s and ITS2, but different annealing temp (sorry machine is running so I can't check ITS2 temp)
__________________
Microbial ecologist, running a sequencing core. I have lots of strong opinions on how to survey communities, pretty sure some are even correct.
thermophile is offline   Reply With Quote
Old 01-24-2017, 03:56 AM   #14
shreve697
Junior Member
 
Location: Pennsylvania

Join Date: Feb 2016
Posts: 6
Default

I just wanted to say thanks again for all of the advice. In case anyone stumbles on this post in the future, I wanted to wrap it up by stating my final conditions. I ended up using 0.25 ng/uL final template concentration for the 515f/806r primer set with 25 cycles, 56.5 C annealing temperature, and a 20 second annealing time. For ITS3/ITS4 primer set, I used 1 ng/uL final template concentration with 30 cycles, 59.5 C annealing temperature, and a 20 second annealing time. Gels looked good, I'll be submitting for second round PCR and sequencing this week!

Thanks again!

One additional question:
During one of the final rounds of PCR, I forgot to tighten down the heated lid of the thermocycler. The heat was still turned on (105C), but the lid was not in contact with the PCR tube caps. After the run, there did not appear to be any condensation on the upper part of the tube, but maybe a few more microdroplets near the top of the "fog line" that normally occurs where the tube is in contact with the block as it's cooled to 4C. Could this introduce a significant amount of error into community analysis? Should I re-run these? I hesitate, because it was a total of 132 reactions run in this manner and a significant amount of labor and materials involved. I would appreciate any opinions on the matter. Thanks you.
shreve697 is offline   Reply With Quote
Reply

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off




All times are GMT -8. The time now is 12:45 AM.


Powered by vBulletin® Version 3.8.9
Copyright ©2000 - 2017, vBulletin Solutions, Inc.
Single Sign On provided by vBSSO