Problematic ddRAD library post-size selection bioanalyzer trace

[nextgenNOOB]

Hi everyone,

I'm preparing a ddRAD-seq plate for Illumina HiSeq 4000 sequencing and I'd like to know if anyone has seen similar Bioanalyzer traces post-size selection with the BluePippin. The first image below shows my traces post SS (left) and pre-SS (right). In our lab, we clean up the DNA before PCR but after digestion/ligation, and size selection is performed post-PCR.

I used a fairly broad size selection window of 300-500 bp. The Bioanalyzer trace of my library before size selection shows a fairly even fragment size distribution, though with a slight bias towards larger fragment sizes. However, after size selection, the Bioanalyzer trace shows a sharp peak rather than a plateau between 300-500 bp. For comparison, I also included a photo of a Bioanalyzer trace of a different post size-selection ddRAD library a lab mate produced several months ago.

Has anyone experienced a similar issue before? I'm assuming sequencing will produce sufficient reads from this plate, but I am anticipating more plates of these species in the future and I'd like to combine datasets with similar read distributions (i.e. ideally I would have a plateau distribution from 300-500 bp post-size selection).

Thank you!

https://imgur.com/deOdQK7

https://imgur.com/ivfA6cM

[nucacidhunter]

It looks fine with short fragment tail which could be due to overloading or the cassette type. The reason for profile difference could be size distribution of library fragments which for the present is pretty even and library peak is median of selection range. The old library profile must have been different in size selection range and have had some bands from repetitive region of genome so its profile differs.

I think 200 bp window is wide and would require more sequencing to obtain good coverage across tags.

[nextgenNOOB]

Thank you for the reply but I'm not convinced that any of the possibilities you mentioned could explain what I'm seeing. We've done several plates now with the exact same protocol, cassettes, reagents, etc. Our former plates have been made up of species from the same family and even from the same genus as the one I'm working with now, so I would expect any repetitiveness of the genome to be seen in my samples too.

Before size selection, my trace looked just like the traces of other plates we've had pre-SS, it's only after size selection that they look different (though we're using the same selection range). We are getting the same avg bp size, it's just the mine fragment distribution is peaking whereas everyone else who has been using this range has been seeing a plateau. I've overloaded a bioanalyzer chip before and it had a different effect from what I'm seeing now (but I only added 1.9 ng of DNA to produce the trace in question).

Also (bc now I'm curious!), our lab has sequenced successfully with a 200 bp window--have you found that you get significantly more reads with a smaller bp range?

[nucacidhunter]

By overloading I meant Pippin cassette not the BA which has resulted in distinct left shoulder.

From your statement it seems that size selection is the only issue. I would suggest to perform another size selection (you would have some PCR amplicons or cleaned digestion-ligation reaction for fresh PCR) to rule out faulty Pippin cassette lane.

Size selection window depends on how many polymorphic loci is required for experiment. Generally with larger window one gets more RAD-tags and deeper sequencing is required to have sufficient coverage. Other issue with large size selection window is uneven coverage as it would cause preferential amplification and sequencing of smaller tags.