RAD-seq on MiSeq?

[web]

I've never done RAD-seq before but I have the opportunity to do a MiSeq run. RAD-seq data would be most useful to me at this point (no current genomic information on my taxa), but I've not seen anyone use MiSeq for this purpose. I assume it's due to lack of sequence depth, but I wonder if there are other concerns as well. Has anyone used a MiSeq for RAD-seq? Are there other reasons why this would be a wasted effort?

If I push forward with RAD-seq analysis using the MiSeq, should I reduce the number of individuals or or modify protocols to somehow significantly reduce the number of fragments?

Any suggestions and/or discussion would be helpful. Thanks!

[GenoMax]

This is a past thread about a similar question: http://seqanswers.com/forums/showthread.php?t=40803

SNPsaurus regularly participates in the forum and may have additional information available later today.

[SNPsaurus]

Running RAD-Seq on a MiSeq is done with no problems. If you have the option to use a HiSeq then you'll get better results, since the cost of a MiSeq run is about the same but you end up with longer reads instead of higher depth.

What's your species? You probably should think about multiplexing fewer than 10 samples using an infrequent cutter. In my lab we sometimes reduce the number of loci by making the RAD library and digesting with a 4-cutter, although this makes the fragments more like a ddRAD library so isn't recommended if you have a highly diverse set of samples.

I would think hard if the convenience of running samples locally outweighs the much lower multiplexing. Most facilities are set up to receive outside samples. The University of Oregon facility runs lots of external RAD libraries, for instance (disclosure, the company SNPsaurus uses the facility and as an academic I am on the facility advisory board), with a lane of 100 bp costing $1901, which is probably what your MiSeq run costs. Also, Allseq or Genohub will connect you with facilities looking to run anyone's samples. But, if the MiSeq is the way you want to go, then just make sure your fragment lengths are appropriate for the read length.

[kbusseq]

Hi SNPsaurus - I am a tech at a facility that has recently started running more external RAD libraries and we've been having a really hard time determining what loading concentration to use on the HiSeq. We just ran a chip with 8 RAD libraries and used the same loading concentration for all of them (based on a previous run for one of the user's samples) and had cluster densities ranging from 150K/mm2 to 950K/mm2 per side. Has your facility been having similar problems? Are there any aspects of the libraries that need to be taken into consideration when determining the loading concentration? I know average fragment size can have an effect, but other than that I'm at a loss.

Thanks!

[manjari.deshmukh]

Hi,

I am trying to produce phylogenetic analysis using RAD-seq. Anyone have any idea on how to go about it