[GenoMax]

Perhaps you can get a HiSeq2000 now (instead of the 1000 if it has not shipped) then do the 2500 upgrade after the middle of year. That may be a conservative option (if your budget permits this).

[epistatic]

The upgrade isn't as exciting as it first appeared but could be useful. I'm still figuring out if worth $50K to upgrade to make our HiSeqs fit in between our MiSeq and HiSeq. Run time of a MiSeq with the per lane output of a HiSeq.

HiSeq 2500
Speed is due to the ability to run a flow cell with only two lanes. The quick runs are due to the lack of needing to image the whole flow cell and improvements to imaging (?). With a flow cell that only has two lanes, I can see the price per lane need to increase slightly versus running all eight on the standard HiSeq. The SBS usage will scale, but I doubt they discount a 2-lane flow cell to be equivalent to 1/4 price of the current flow cells.

With Illumina's HiSeq specs combining the two flow cell positions, it appears the 120 Gb comes from running two two-lane flow cells with 60 Gb each. With PE-100, that is ~150M reads *2 per lane (less than we get now). If the 120 Gb comes from a single two-lane flow cell with ~300M reads * 2 and could run in 27 hours, that would be something.

[simonandrews]

I doubt they discount a 2-lane flow cell to be equivalent to 1/4 price of the current flow cells.

The cynic in me wonders what the QC failure rate for new flowcells is, and whether they're selling on some of the partial failures as 2 lane flowcells

From my point of view, if that's what the upgrade actually is I'm not too worried about missing it. Now if they could offer the ability to run arbitrary numbers of lanes on any flowcell so we could use partial flowcells routinely and come back later to use up the lanes we didn't use the first time, then I'm really interested, but just having a 2 lane flowcell isn't quite as exciting.

[ScottC]

I doubt they discount a 2-lane flow cell to be equivalent to 1/4 price of the current flow cells.

No, they won't. Illumina has already said that the cost per base will be higher for the lower-yield instrument runs.

Scott.

[epistatic]

I also just learned Illumina pricing will increase 8-10% March 1. The need for speed would be good for small pilot projects people bring us at the last minute for grant applications but I'm not sure the quicker turn around time will justify making back $50K for the upgrade. Most people need weeks to months to analyze their exome and genome data.

[simonandrews]

Just heard from our rep. Another interesting tidbit is no more need for a c-bot. Cluster generation is done on the system a la the miSeq with the 2500.

The on-flowcell cluster generation only works for the 2-lane flowcell. You still need a cbot to prepare a full run.

[simonandrews]

Any hints as to whether there might be an upgrade to the HiSeq 1000 to provide the sorts of improvements seen in the 2500?

To answer my own question, there is indeed an upgrade from a 1000 to a 1500 available.

[kmcarr]

To answer my own question, there is indeed an upgrade from a 1000 to a 1500 available.

So you're saying there will be a 1 flowcell version of the 2500?

[simonandrews]

So you're saying there will be a 1 flowcell version of the 2500?

I don't know if this will be sold as a separate product, but we've been told that we can apply the 2000 > 2500 update to the 1000 we're about to get and that this is a supported upgrade path. I guess they're all based on the same platform so it makes sense.