[aeonsim]

I agree this is a stupid restrictions. However, is there a lot of drive to do huge scale, medium depth, WGS on a species other than human? How many 30x mouse genomes can one need?

You aren’t going to get a de novo genome with 30x at a single fragment size, at least not one worth much. And the de novo sequencing must make up a tiny fraction of sequencing projects in the world. So, if Illumina is going to have some 5 different machines now, what the problem in having one machine dedicated to large scale human resequencing studies? Maybe the implementation is wrong, but I think the general idea is right, especially from a marketing perspective.

At least with Cows there are multiple groups sequencing as many animals as they can afford to get a better understanding of variation within and between breeds. Even with the previous prices the groups were sequencing between 140-700 animals each (plus a community 1000 Bull genomes project), if the price went down we'd sequence more and apply for more funding (both industry & public) to get a decent understanding of the different breeds and the range of variation in the population which is critical for animal breeding etc

[ymc]

There is still one thing I don't quite understand about HiSeqX. That is: do you need to buy it in 10-pack? It appears to me Table 1 in their data sheet shows the stat for one box. It doesn't seem to me you need 10 boxes to run together to archieve that throughput.

Is it because if you buy it in 10-pack, then you get a significant hardware discount such that the lower instrument depreciation cost can push the cost of 30x genome to be under $1,000?

[AllSeq]

There is still one thing I don't quite understand about HiSeqX. That is: do you need to buy it in 10-pack? It appears to me Table 1 in their data sheet shows the stat for one box. It doesn't seem to me you need 10 boxes to run together to archieve that throughput.

Is it because if you buy it in 10-pack, then you get a significant hardware discount such that the lower instrument depreciation cost can push the cost of 30x genome to be under $1,000?

The restriction is that the minimum purchase is 10 machines at a time (but they don't need to be purchased only in 10s - someone had an initial order of 14). There has been nothing to indicate that the instruments operate together - it's just a minimum purchase defined by Illumina (essentially for pure marketing purposes - if you can't afford 10, you aren't in the right customer segment for this machine). Same logic behind restricting it to running human samples - pure marketing choice.

[ymc]

Really, this is a bit of a giant FU to Complete Genomics and, uh, BGI.

Yeah, this HiSeqX has 10x throughput over HiSeq 2500 SBS v3. Obviously that's a very big FU for BGI indeed.

But I heard that they have more management issues than technical issues.

Anyway, if Dr Yang can convince the right commies again to buy a bunch of HiSeqX Ten, he can still be in the game. The NYT news today said China printed more money than the US.

[ymc]

The restriction is that the minimum purchase is 10 machines at a time (but they don't need to be purchased only in 10s - someone had an initial order of 14). There has been nothing to indicate that the instruments operate together - it's just a minimum purchase defined by Illumina (essentially for pure marketing purposes - if you can't afford 10, you aren't in the right customer segment for this machine). Same logic behind restricting it to running human samples - pure marketing choice.

That makes sense now. I suppose that price might also include the salary of a local support person. I know they hired several support guys in Hong Kong when BGI bought the 128 machines.

[TonyBrooks]

The rep at PAG said they planned on selling 4 "X boxes" in 2014--and had already sold 3 30 minutes after the announcement.

The NextSeq500 is clearly aimed at people considering a Proton. Same price, same emphasis on speed, better stats right now. I don't like that the only single-end mode is 75 bp. Lots of people like the 150 bp read on the 2500 Rapid. And the other downside is that all 4 lanes (100M reads each) get fed the same library, so doesn't improve flexibility in flow of projects at a facility. But a nice mid-machine.

There is a 150 cycle kit. It's marketed as 75 paired end, but if 150 read length is supported, I'd we surprised if you're not able to do a single 150bp read (just like you can use the MiSeq 150 v3 to do 150 single read or 75bp paired end).

[pmiguel]

What I'd like is some more detailed explanation of the 2-dye system the NextSeq 500 is using. I assume they provide C & A tagged with Red, T & A tagged with green & G untagged, thus C should be pure Red, T pure green and A ~ a 50:50 mix of green & red (orange) and G then undyed...

So if you have 5 Gs at the beginning of a template, there will be no cluster found?

Would not a bubble in the flowcell result in all those clusters being called "G"? (At least for the bottom surface.)

[TonyBrooks]

So if you have 5 Gs at the beginning of a template, there will be no cluster found?

Would not a bubble in the flowcell result in all those clusters being called "G"? (At least for the bottom surface.)

Surely anything beginning with 5 G's won't be registered as a cluster. So bubbles wouldn't contain any clusters either - so you don't get any data. I would expect that there is some algorithm there to detect when something is a G and when it's an intermittent bubble. G's would be localised areas with no signal, bubbles would cover much larger areas.

You'll end up losing less than 0.1% of your clusters assuming total randomness (1 / 4^5) but I guess that's a hit worth taking to effectively half the imaging time.

One thing I'd like to know is whether NextSeq employs the empirical phasing calculations used on the MiSeq, or whether that's too computationally expensive.

[Chipper]

Do they really need the five first bases to identify clusters on an ordered flowcell?

[GenoMax]

Do they really need the five first bases to identify clusters on an ordered flowcell?

AFA I can tell NextSeq 500 does not use ordered flowcells and the two color chemistry is only for that instrument.

[AllSeq]

AFA I can tell NextSeq 500 does not use ordered flowcells and the two color chemistry is only for that instrument.

Yeah, the ordered flow cells appear to only be associated with the new HiSeq X machine (at least for now).

[GenoMax]

Illumina appears to be referring to ordered flowcells as "patterned" flowcells. We need to get used to that term.

[Chipper]

Ok, thought the post was about the X ten. Why would they not use the two color system for it?

[AllSeq]

Not sure where you got that from or why they would not include the new flow system, chemistry and optics in a machine that was designed to reduce cost and increase throughput. This is from http://www.illumina.com/systems/hise...ng-system.ilmn :

Building on the proven performance of Illumina SBS technology, HiSeq X Ten utilizes a number of advanced design features to generate massive throughput. Patterned flow cells, which contain billions of nanowells at fixed locations, combined with a new clustering chemistry deliver a significant increase in data density (6 billion clusters per run). Using state-of-the art optics and faster chemistry, HiSeq X Ten can process sequencing flow cells more quickly than ever before – generating a 10x increase in daily throughput when compared to current HiSeq® 2500 performance.

I think we're saying the same thing. I agree that the HiSeq X will use the new patterned flow cells. My supposition is that these patterned flow cells will ONLY be for the HiSeq X and it is based on two things:

1) They've only mentioned them for use on the HiSeq X
2) In conversations with people from Illumina (prior to Tuesday's announcement), they said that the patterned flow cells would be restricted to the human whole genomes.

I don't think there is any technical reason to limit these new flow cells to the HiSeq X. I think it's just a marketing decision.

[AllSeq]

Our take on the "$1000" genome

On our blog we've given our take on Illumina's "$1000" genome and why it still isn't available for most people.