Optimizing on NextSeq

[MCA]

Howdy,

I work for a clinical genomics company, we have a NextSeq 500 and have been trying to squeeze the most reads per sample out of it.

What I've been running:
Multiplexed 24 Human DNA samples
TruSight One enrichment panel, 8bp barcodes, paired end
NextSeq MidOutput Flowcell, run as 2x 150 cycle

We've been consistently been getting:
60Gb data
180million sequences
~70-80% Passing filter
~70-80% Q30
Samples have average 80x coverage of the panel

I've been able to get that high by loading a much higher pM concentration on the flowcell, and finding out at which concentration it begins to overcluster and filter out data.

I'd like to know, is this normal? I assume other labs are trying to maxmize data, dose anyone have similar (or better) results? Illumina seemed tight lipped about our results, neither condemning or condoning us overclustering. We are planning on switching to a HiSeq 3000, would we be able to go beyond (at our own risk) the recommended specs?

I heard that the HiSeq has patterned flowcells, would this set a ceiling to prevent clustering too high?

Also, if you have any NextSeq questions feel free to ask!

-Plunk

[ScottC]

In our experience, the NextSeq is by far the most tolerant of the Illumina systems in terms of pushing the cluster numbers (second after that is the old GA/II/IIx), although the patterned flowcells of the 3000 might change that. We get similar figures to those you posted on our NextSeq.