PhD-students-driven paper is published in Molecular Ecology Resources. Read it to learn more about chimeras in RAD-seq.
The wealth of genotyping that we have done in the last 5 years (and by “we”, I mean Marisa and Rohan), coupled with the upgraded quaddRAD genotyping protocol, enabled us to characterise and quantify the various chimeric sequences in a sequencing output and assess their impact on the interpretation of the results. We knew we would be able to do this once Marisa switched to the four-adapter protocol by Franchini et al., and she and a fellow PhD student at Huddersfield, Bernardo, went away with the analysis. Rohan’s project delivered another few hundreds of samples genotyped with the same adapters and protocol, and in two different PCR setups. We had all the pieces for a really high-powered study. So, what did we find?
The bad news is that chimeras are there and most of the protocols do not allow for their identification at all, particularly if you put several multiplexed groups of samples on a single lane (as you most likely do). The good news is that if you maintain good quality standards in your wet-lab and computational pipeline (high molecular weight DNA, high coverage filters etc.), impact of the chimeras is minimal. We also have found that you should probably not attempt barcode rescue beyond 2 mismatches, as above that level you may be mainly creating chimeric rather than recovering actual reads.
Enjoy the paper and congratulations to all involved!