[krobison]

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.

For a lab that already has a HiSeq or IIx, you are probably correct -- the cost per base (even amortizing in the machine) will be significantly higher on the MiSeq.

But, for someone not yet in that game, it is a very interesting machine -- particularly if $125K makes people furrow their brow but $500+K causes them to stop taking you seriously.

ALSO, note the run times. For some applications, being able to go from sample DNA to data in a matter of hours or just over a day (depending on the read length & amount of data you need) is critical. I suspect that even some labs well practiced with the "big iron" will consider a MiSeq if they have applications requiring fast turnaround or to do small method-development projects.

Also, any word about machine cost? I get the feeling they are trying to compete with IonTorrent.

$125K, for a complete package. Very similar to what is claimed for Ion Torrent once you throw in necessary accessories. And not just Ion Torrent, but 454 Jr (and to some degree PacBio, given PacBio was going to offer fast runs cheaply -- ignoring the amortization cost of the instrument).

[genlyai]

Exactly, we don't have a HiSeq, we won't be buying a HiSeq. We send our stuff out because we don't have that much of it.

But we do make some unusual libraries from time to time. From time to time we try to run a sketchy sample. It is really annoying to wait 4-6 weeks to find out if it worked. (It would even be annoying to wait 5-10 days for the big machine to run if we had it on hand.)

[ymc]

Can MiSeq be used for full exome sequencing in one run???

[krobison]

Illumina is claiming just under a Gigabase in the 2x150 mode. For a 50Mb human exome capture, that would be only 20X coverage, which seems like a bit of an undershoot.

On the other hand, with a 5Mb custom capture design it would be 200X average coverage, which would seem doable (what are folks shooting for these days?).

[ymc]

Illumina is claiming just under a Gigabase in the 2x150 mode. For a 50Mb human exome capture, that would be only 20X coverage, which seems like a bit of an undershoot.

On the other hand, with a 5Mb custom capture design it would be 200X average coverage, which would seem doable (what are folks shooting for these days?).

Thanks for your reply. How many x of coverage is required to reach the genotyping accuracy of the microarray chips?

[Vinz]

Sounds like a great instrument!
Whats your bet on read length? Do you think Illumina is at the limits of the chemistry with the 2x150bp?

[avilella]

Has anyone done the math on the price per base one would get with a 2x150bp run on the MiSeq and the HiSeq, taking into account library prep costs?

People working medium-sized genomes (10-100 MBp) sometimes spend a lot of time and effort getting clean clonal DNA, so doing a try on the MiSeq would be great for them, but maybe they can wait to do 1 lane on the HiSeq if there is not much price difference...

[genlyai]

The price per base difference will vastly favor HiSeq, but I'm not sure that's the metric you would care about if you want to quickly test the quality of your library & DNA. Wouldn't you just care about turn-around time and cost per run?

[vebaev]

Hi,
our lab have used only sequencing services for mRNA-seq and smallRNAs seq with Illumina GA (for plant samples).
We are quite interesting to get MiSeq since we do not have sequencer, is there any limitations that MiSeq have that we have to know before buying it?

Thanks

[vebaev]

Since the covarage of MiSeq will be smaller than GA is it possible to use it for sequencing of transcriptomes (mRNA-seq, plant species in my case)? If not whole transcriptome is it possible to catch at least most abundant transcripts?

The same for genome sequencing? is it possible to catch part of the genome with Miseq (again plant genome) and If needed to make more runs for more covarage of the genome?


Thanks!

[james hadfield]

avilella & genlyai: The price per base is much higher on MiSeq than HiSeq but the ability to get a result in 1 day vs 8 i imporatnat in many areas. I also find that users who wnat a long run often wait longer than usual as we do fewer long runs compared to SE36 for instance. In this case the 1vs8 days might be more like 1vs30 in reality. Almost certainly worth the extra cost in that instance.

[james hadfield]

Vinz: The MiSeq completes a PE100 run in 1 day compared to 8 on HiSeq and 12 on GAIIx. If you look at intensities and then error profiles for runs they drop off at the end. Much of this is due to the time the reagents sit on the instrument and Illumina only support ten days.
MiSeq runs fast, very fast. I am hoping that this has an impact on quality and would not be surprised to see users publishing 200, 300 and even higher read lengths. Lets wait and see when people get hold of them.

[nickloman]

Wow, that's an interesting thought!

[Vinz]

Thanks James for your thoughts! I hope you are right - a little above 200 bases would be just perfect for our application.

[ECO]

Vinz: The MiSeq completes a PE100 run in 1 day compared to 8 on HiSeq and 12 on GAIIx. If you look at intensities and then error profiles for runs they drop off at the end. Much of this is due to the time the reagents sit on the instrument and Illumina only support ten days.
MiSeq runs fast, very fast. I am hoping that this has an impact on quality and would not be surprised to see users publishing 200, 300 and even higher read lengths. Lets wait and see when people get hold of them.

I would guess that the decay of phasing and overall efficiencies would be a bigger limitation than reagent stability. Even at >99% efficiency, for 100++ sequential chemical reactions the intensities (sum of successful incorporations) and signal-to-noise (ratio of correct incorp to all three incorrect) would drop accordingly.