ezRAD and other RADseq

[blair.chr]

Hi all,

I just came across a paper describing a new RAD protocol termed ezRAD (https://peerj.com/articles/203/). In short, you simply digest your DNA with restriction enzymes and then use Illumina's standard TruSeq library preparation kit to make your libraries. Has anyone used this method yet?

I just started a tenure-track position at a teaching college and have limited space, time, and funds to create my own RAD libraries. Floragenex would charge at least $79/sample, which seems pricey to me. I have experience creating ddRAD libraries, but am not sure if this would be feasible at my new institution.

I would welcome suggestions regarding optimal ways to conduct RAD-based population genomic studies at primarily teaching institutions.

All the best,
Chris

[SNPsaurus]

I think the ddRAD protocol is simpler or the same as ezRAD if you are doing it yourself. The strength of ezRAD is that the material can be sent to a core facility for library prep, but it would be difficult to find a facility willing to make very cheap libraries. For instance, Texas A&M Agrilife, which does great work, charges $175/sample for TruSeq DNA. This cost is negligible if it is going for full-genome sequencing, but too high for genotyping!

In my academic side, I teach an undergraduate genomics course where students do RAD-seq on samples of interest and do a little analysis. They come up with fun things like comparing yeasts in Belgian beers, looking for traces of insects in different campus eateries, comparing the mouth microbiome of dog owners that kiss their dogs to those that don't, etc. Could you have a course where you enlist students to work on your samples? Another faculty member here has students work on a saturation mutagenesis and it works quite well. Although I am a fan of RAD-Seq and nextRAD since my lab developed them, ddRAD does not need a sonicator so a small set of dual-index adapters could handle a large population (but is admittedly still pricey). On the other hand, if you make multiple libraries some exact method of size selection is needed (this appears also to be true of ezRAD and is born out by the poor performance of SNPs found across multiple individually-prepared libraries in the paper).

In my experience, if you start with lots of high-quality DNA, every protocol is easy and undergrads have had no trouble even with the longer RAD-Seq protocol. But I realize that it is one thing for students to come into a lab and do a protocol with guidance from grad students that spend all their time doing that protocol and trying to do it starting from scratch. But tying it to a course lets you apply academic funds to your research, and it is hard to beat that!

Hmm, it seems that by the end of all this I don't really have a great solution for you. Floragenex charges what it does because it builds in all sorts of extras to help ensure success. A lab will spend 6 months working out a protocol, ignoring the labor cost involved, have the run fail and chalk it up as a learning experience. But if everything works it is possible to make the libraries yourself and simplified protocols like ddRAD are quite do-able in a short amount of time with a minimum of investment.

[blair.chr]

Thanks for the prompt reply. Yes, it may be possible to solicit the help of undergrads to create ddRAD libraries, but I am a bit hesitant. Also, some problem I have run into with developing your own ddRAD libraries include 1) size selection (i.e. finding a Pippen-prep), and 2), finding a sequencing core that will sequence self-made libraries. However, I agree that costs are modest and mostly lie in the Ampure beads and barcoded adapters.

I guess I will have to weight the pros and cons of purchasing equipment and reagents and creating a ddRAD library in-house versus outsourcing to Floragenex for single digest RAD.

-Chris

[SNPsaurus]

The problems with ddRAD and missing data come from making multiple libraries and performing separate size selections for each. If you have a population of 96 and do a single size selection on the entire library, and plan a combination of restriction enzymes so that a relatively broad size selection is OK, then you can do without the Pippin Prep and just do a gel cut for size selection without too many issues.

The University of Oregon facility sequences plenty of RAD-type libraries without doing the prep as well. The new Illumina software upgrade also makes sequencing reduced-complexity libraries much better as well so it is a good time to do genotyping by sequencing! But if I had a limited budget and wanted to reduce the risk in doing a project, I would outsource almost every time. If I had money and time to tolerate failures, I'd do it myself.