Realistic output for HiSeq 4000

[kmcarr]

The official spec for PF Clusters on the HiSeq 3000/4000 is 270-312M per lane. It seems that it is pretty easy to exceed this specification but I would like to ask regular operators of a 3000/4000 what PF cluster counts do you reliably get per lane? Is there a difference depending on library type?

Thanks.

[misterc]

It does depend on library type as well as insert size. I usually get 350-400M reads passing filter per FC.

[luc]

It does depend on library type as well as insert size. I usually get 350-400M reads passing filter per FC.

Agreed. Long insert libraries (library molecules longer than 550 bp) will reduce yields and quality scores.

[GenoMax]

Yield is not an issue as long as libraries have the right insert sizes. Illumina guidance is around 75% PF clusters for a good run.

There are optical duplicates but they appear to be randomly distributed when checked recently (n=1).

[dpryan]

We usually get 300-400 million passing filter per lane on a HiSeq3000. "Optical" duplicates are typically below 10%, except with weird library types. Often the higher the read yield the lower the "optical duplicate" rate.