How many PCR cycles should I use to create an amplicon library?

One of my side projects deals with the gut microbiota of Atlantic Cod (Gadus morhua). In 2014 we published a pilot experiment on this where we sequenced the 16S rRNA to get a taxonomical overview of the microbial gut content of 11 Cod specimens caught in the Oslofjord, Norway.

The Cod in the pilot experiment were wild fish and we had not really an idea what we would find. So it was a bit of a surprise that these 11 fish had only a few species of bacteria in common in their gut. Most species belonged to the order Vibrionales and Bacteriodales.

Now a PhD student in our lab is gearing up to repeat this experiment with more fish and more locations. He has finished the DNA extractions and is now preparing for making an amplicon library. Since the samples contain various amounts of DNA content and quality, he tries to optimize the PCR reactions so that all samples have enough sequences after the amplicon sequencing. At this point we got into a discussion on how many PCR cycles one should use to have a good balance between enough sequence templates and not too many PCR artifacts that you need to remove afterwards.

Since I was not entirely satisfied with my own answers (being strict on using a low amount of PCR cycles) I posted a tweet with my question online and I made a little story of it using Storify.

How many PCR cycles should I use to create an amplicon library?// The result is below.

How many PCR cycles should I use to create an amplicon library?

The answers I got at twitter for my question on PCR cycles

  1. This morning I posted the following question on twitter:
  2. Love to hear some arguments why I should not use 35 cycles for my pcr amplicon sequencing experimemt. Chimera’s, pcr errors, anything else?
  3. I then got into a nice little conversation with Mike Cox:
  4. @Thomieh if you’ve got enough template/target to use fewer, then not justified?
  5. @MikeyJ thanks, but what if some samples have enough and other do not? Is that a problem afterward during analyis?
  6. @Thomieh that can have a big impact, I go for the higher number cycles that most will work at. I avoid higher than 35.
  7. @Thomieh similarly biased is the compromise, but you end up tolerating some drop outs depending on sample type
  8. @Thomieh a 16S qPCR can really help to judge whether there’s nothing worth chasing in a sample or you should up cycle nunbers
  9. @MikeyJ okay, I had not considered the q-pcr yet, but that is surely a solid way of testing samples. I’ll try that.
  10. @Thomieh we only introduced it latterly and it’s really helpful for lots of interpretation e.g. weird bug in one sample + low qPCR = contam?
  11. @Thomieh also if you have longitudinal samples, you have proxy for activity – increase in proportion + increase in qPCR between two samples
  12. @MikeyJ that would make it even better. Do you test all samples with the q-pcr? Or just those that are dubious on a gel?
  13. @Thomieh all of them, sometimes we sequence products that you can’t see on a gel. It’s not too expensive, especially in plate format
  14. @MikeyJ but if the input is so little, is that still reflected by the data output of such small samples. I am a little worried there.
  15. @Thomieh yes, you’re right to be! Usually they end up being thrown out by rarefaction as too few reads. Depends on experimental design
  16. @Thomieh if they’re super precious unique samples and longitudinal study, we go for it and keep it in mind during analysis
  17. @Thomieh if there are large numbers in categories to compare we don’t force it. Controls often low biomass and still important to seq tho.
  18. @MikeyJ ah that makes sense. I think you answered a few of my worries here, and we will try the q-pcr for some more control of the input.
  19. @Thomieh no problem, I think there’s no firm rules when it comes to this, it’s all about reducing and justifying bias.
  20. @MikeyJ Yes that is also what I think. Thanks for the input. I will put the tweets in a little blog to keep it. Okay?
  21. Feel free to comment on this twitter story, here or directly to me at twitter. @Thomieh


Any comments, additions are welcome.

About Thomas Haverkamp

A microbial ecologist, an amateur photographer and a proud father a wonderful girl.
This entry was posted in High-throughput Sequencing, Microbes and tagged , , , , , , , , , , , . Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s