HiSeq rapid run overclustering

[anna_m]

Hello,
On the Illumina website, it says that the optical raw cluster density for PhiX runs is 850K-1000K clusters/mm2. But, what is the functional range of the instrument in this mode? The GC bias in over clustered situations using the Rapid chemistry has not been tested and I was wondering if any other labs had advice about this.

Thanks,
Anna.

[ECO]

We typically run 1100-1200 (still at ~90% PF), but with the newest software on originally built 2500's. 2000's that were upgraded to 2500's are less capable of handling >1000k. Things go south rapidly above 1200.

[GenoMax]

Are you asking if you can deliberately overcluster beyond that spec? You may be able to (to some extent) but we have found this to be library dependent. It may work with one library but may not with another. In any case you should stay under 1100 or so with upgraded 2000's like ECO said above.

[sajoshi]

Yes that is correct, ideally is 850 to 1000. I agree with Genomax that clustering also depends on the sample type, library etc. Libraries do perform differently, so clustering will differ. Do not increase loading conc. to increase clustering. You don't want it to over cluster because it may affect % PF. Stay under 1000 for clustering.

[misterc]

There are 3 different cameras on HiSeq 2000's in the field. The oldest of them will really struggle at densities above 1,000K/mm2. The other two camera models will fare better with higher densities. Manufactured HiSeq 2500's all have the newest of the camera models and will be more capable of resolving densities of 1,200-1,400K/mm2.