[csc]

Anyone want to bet on how high Illumina are going to predict HiSeq will go by the end of 2011?

My pick in the sweepstake is 2Tb, it is HiSeq 2000 after all!

Are we talking internally at Illumina or commercial release? It looks like this spring it will be 600Gb officially (using 2x100bp) but it seems like all they have to do later in the year to get it to 1Tb+ is officially release the 2x150bp chemistry. Easy upgrade for them. I'm not totally sure we'll see 2Tb commercially this year, not because it isn't possible, but because Illumina may not have a reason to push it that high unless ABI has something up their sleeve.

[james hadfield]

In my post above I was not clear about what I thought MiSeq might be in comparison to GA or HiSeq capacity. I meant 10% of a lane "The imaging for ten ‘tiles’ would be about 30sec.".

If ten tiles is correct then doubling this would only add about two hours to a run to double data volume. Maybe Illumina will make this bit flexible as well. Then users could dial in run time and data volume.

BTW, many people have already done this on GA or HiSeq.

[beelu]

BTW, many people have already done this on GA or HiSeq.

Hi James, can you elaborate on this? Do you mean its possible to run GA in shorter period of time with less output data? (less reagent?)

Thanks.

Beelu

[nickloman]

A little more info from the UK rep here

http://pathogenomics.bham.ac.uk/blog/2011/01/a-few-more-miseq-nuggets/

[HESmith]

Hi James, can you elaborate on this? Do you mean its possible to run GA in shorter period of time with less output data? (less reagent?)

Thanks.

Beelu

Run times can be shortened by imaging only a subset of tiles (or lanes).

[james hadfield]

I was wondering if it might be possible to get even more fomr this box of tricks by daisychaining some of the internals.

Right now the imaging module is the most expensive bit in a GA or HiSeq. This is probably the same for MiSeq. As imaging takes 30sec compared to five mnutes for chemistry why not simply have six fluidics modules and put six 10 tile single lane flowcells on, or just one. All would be imaged by the same module just like HiSeq.

MiSeq 2000. You heard it here first ;-)

James.

[genlyai]

So have we heard anything about $/run and how that compares with Ion Torrent? If that's comparable, this seems like it's going to crush Ion, mainly because of Nextera prep.

[nickloman]

$400 - $750 per run apparently.

[genlyai]

Thanks, Nick. I see now that that was on your first blog post. Is that price inclusive of library prep?

If so, MiSeq is winning on every metric -- not least that it will use the exact same libraries as the big machine. As someone else said, it's not really in our interest for Illumina to leave its competition in the dust, but that's what seems likely here.

Assuming the machine meets these specs. I guess we do know by now that that's a big assumption.

[GW_OK]

Anyone want to bet on how high Illumina are going to predict HiSeq will go by the end of 2011?

My pick in the sweepstake is 2Tb, it is HiSeq 2000 after all!

I'd bet a dollar it'll be 1Tbp by summertime, so you might be right...

[nickloman]

Thanks, Nick. I see now that that was on your first blog post. Is that price inclusive of library prep?

Not sure, perhaps not - but a single library prep with Nextera is pretty cheap I believe (? £30 per sample).

[nickloman]

I'd bet a dollar it'll be 1Tbp by summertime, so you might be right...

They did a 1.17Tb run already at Illumina HQ, I believe!!

What an amazing time this is!

[GW_OK]

They did a 1.17Tb run already at Illumina HQ, I believe!!

What an amazing time this is!

Yes, but it's not supported and in our hands. It should be by summer.

[GERALD]

It sounds like the same reagents on a lower capacity machine. I think many labs would rather not go to the trouble, especially when you consider that the amount of labor is the same. Also, any word about machine cost? I get the feeling they are trying to compete with IonTorrent.

[genlyai]

It sounds like the same reagents on a lower capacity machine. I think many labs would rather not go to the trouble, especially when you consider that the amount of labor is the same. Also, any word about machine cost? I get the feeling they are trying to compete with IonTorrent.

Well, consider the sample prep will supposedly be a matter of a few hrs with a few minutes of hands-on time. Meanwhile the run time, cost per run, and upfront cost are much smaller. There are definitely applications, not least (as Keith Robison always says) validating libraries before they go to the big machine.