NextSeq data bias

[kshazand]

Hi,
We notice an unexpected variation of the estimated library size as a function of the number of reads, even after downsampling with NextSeq data. This is not observed with MiSeq-generated datasets. All analyses were done on Picard.

Did anybody else duplicate this observation? Does this ring a bell?
Thanks for your help,
Kamran

[Brian Bushnell]

If you are using V1 chemistry, that's probably caused by the high error rate. But could you explain a bit more about exactly what the variation was, and your methodology?

[kshazand]

Thanks Brian. We have moved to V2 but the analyzed data were from the V1 period.
Is there an email where I can send you a slide?

[Brian Bushnell]

Thanks Brian. We have moved to V2 but the analyzed data were from the V1 period.
Is there an email where I can send you a slide?

Sure - pm sent.