PCR-free ddRADseq?

[Arbag]

Hi all.

I'm helping out on a RAD-seq project and we're exploring different options to assess population structure for several hundred individuals. I was inspired by the "Rapture" paper (Ali et al Genetics 2016) and wanted to modify the protocol (the pre-sequence capture part). If I understand this paper correctly, they use a single digest+mechanical shearing in combination with a biotinylated 5' barcode-containing adaptor. After pooling+purification with SA beads, they use PCR to add the i5/i7 Illumina adaptors (and another barcode).

I was wondering if it would be a viable idea to instead do a double-enzyme digest and ligate adaptors complementary to either "sticky end" containing both the sample barcodes and i5/i7 adaptor sequences compatible with HiSeq. I would also opt for biotinylated oligos to enrich for RAD-tags. My reasoning is that this would eliminate all PCR clones, leaving you with only RAD-tags from the parent molecule, and you would get more useful reads from more individuals. I recognize that the biotin/SA enrichment will also pull out fragments that were tagged on one end and not the other and the oligos might get obnoxiously long. Are there other major pitfalls I'm not seeing? I appreciate any feedback. Thanks!

[SNPsaurus]

The rapture approach uses the sheared end to detect PCR clones (non-clones are unlikely to have the same break point). With ddRAD and your variant protocol you can't do that, but then you can do PCR-free libraries. I think rapture libraries have reads to spare though, so throwing out the PCR clones doesn't really cost anything and you are essentially creating a problem (no longer being able to detect clones) in order to solve it. The adapter set is likely to be one of the significant costs in the project so for a one-off I'm not sure it is worth it (other than the trials and tribulations and fun of doing something different).