Сollapse of Helicos

[genseq]

http://industry.bnet.com/healthcare/...0-genome-race/
http://www.genomeweb.com/sequencing/...-restructuring

[Calico]

It's a little bit sad to see, and strange since they are one NGS company that actually has a machine on the market. What do you think is the reason for their demise? One thing I have noticed is the lack of marketing on their part, at least if I compare with e.g. PacBio.

[genseq]

The main reason for their demise is monomolecular technology.
Next collapse - PacBio.

[krobison]

A more thoughtful consideration of Helicos might be along these lines

On the business side, they went public too soon. Going public has pluses, but huge minuses too. Lots of pressure to get the product out the door.

Failure of management to enlist a partner with deep pockets -- never easy, but certainly helps keep the engine running.

Instruments were clearly delivered too early to some customers -- having a service provider return your instrument isn't a great start.

The worst problem, and abetted by the issues above, is that they didn't quickly demonstrate enough compelling applications. The Quake genome paper is to me the perfect exemplar of this (yes, it wasn't officially a Helicos paper -- but since he is a founder it inevitably gets assigned to them). It didn't highlight any of the real advantages of the system -- they cooked up cost & labor estimates that nobody took seriously to try to claim some. It did (especially with the recent Lancet follow-up) highlight the limitations of Helicos in terms of accuracy & coverage.

In contrast, suppose they had taken a sample which really showed off their simple sample prep & ability to work with difficult samples. If they had sequenced a person's genome from a small sample or better yet a cancer genome from formalin-fixed clinical tissue, then they'd be doing something ahead of "the big 3" and perhaps in an area where they struggle.

Similarly, while their website has protocols for a number of interesting applications which show the advantages of the system, too few of these have publications. Mass copy number profiling from clinical samples? Why isn't there a publication for that?

Well, because they couldn't do everything at once. This is the challenge of a small company with a hard to characterize technology in a crowded field -- you need to distinguish yourself from the pack AND do so with something that will catch on. Helicos focused on some other apps (RNA-Seq) that didn't quite catch the imagination of the crowd.

Finally, Illumina & ABI & 454 certainly make a lot of money off kits & reagents. Helicos' simplicity of sample prep has the downside of denying them a strong revenue stream in that area.

Helicos isn't dead yet & even if the company dies somebody may pick up the technology.

Finally, writing off PacBio is rather premature. They have a lot of funding & their instrument clearly has some very interesting distiguishing features. Even if it doesn't become a workhorse, it can provide key data that other systems can't (long-range genomics information) and in doing so make obsolete a whole host of physical mapping methods. With the great current interest in structural variation and proposals to map tens of thousands of vertebrate genomes, if PacBio can simply generate noisy reads across 20+Kb fragments, it will have a niche which will support a few dozen machines. There are plenty of companies out there which get by with a few dozen $1M machines in the field -- though that wouldn't support the stratospheric hopes of PacBio investors.

[NextGenSeq]

Thank god I never took a job there. After talking with them it was pretty obvious they wouldn't make it. I doubt PacBio will do much better but they can at least learn from Helicos's mistakes.

[james hadfield]

Krobison makes a really interesting point about revenue streams. Most companies would say they do not make money on instrumentation, it is consumables that drive business. Helicos would make money from sequencing kits, large numbers of users of Illumina are home-brewing kits to save money as well.

However in the case of Helicos, had they been able to show robust RNA-seq from native mRNA sequenced with a reverse-transcriptase, at similar throughput to Illumina I know a lot of people who would have loved to use the system.

[Michael.James.Clark]

I think it's wrong to count out PacBio, especially as a technology. The phasing info alone that can potentially be gained using a PacBio is worth quite a bit.

There is also the potential for direct observation of modified nucleotides using the kinetics data, something that I'm interested to see pan out.

I would not be surprised to see many larger genome centers pick up a PacBio device for the sole purpose of providing scaffolding data to work hand-in-hand with their short read data.

[frozenlyse]

Indeed PacBio have demonstrated on a very small scale their platforms ability to distinguish modified nucleotides without any pretreatment, which makes it a unique technology... But who knows whether it can scale up to useful throughput?

[genseq]

PacBio is outsider in "price race" of "genome race":

[krobison]

Originally Posted by genseq

PacBio is outsider in "price race" of "genome race":

Raw cost of the box is an important consideration, but hardly the be-all, end-all. PacBio. Other costs are the reagents & labor. If you wanted to show something interesting, plot an estimate of the fully loaded cost for some given project over a set lifespan -- say #genomes sequenced over 2 years (assuming 80% uptime) vs. reagent+library prep+instrument cost. That might be informative.

[genseq]

This information will appear in 2 ... 3 years.

[scrosby]

genseq, I like the horse-race! You should update it!

[rskr]

Helicos's physics based approach was a little unrealistic. There is only so much you can do with a giant lens, and last I checked they weren't turning out reads any longer than current short read technology, but with an instrument that was a couple of magnitudes more expensive.

Pacbio with the modified polymerases is much more interesting, and has implications for biotechnology that are fascinating, for example the machine has uses beyond sequencing ACTG, it appears that they will possibly be able to detect any number of DNA modifications methylation just being the first application.

[genseq]

Originally Posted by scrosby

genseq, I like the horse-race! You should update it!

[krobison]

Originally Posted by rskr

Helicos's physics based approach was a little unrealistic. There is only so much you can do with a giant lens, and last I checked they weren't turning out reads any longer than current short read technology, but with an instrument that was a couple of magnitudes more expensive.

Helicos' edge was never in read length. The key advantages were very simple (and cheap) sample prep free from any biases of PCR.

[Apollo704]

Helicos also had a 50 lane set up, and finding enough samples to fill that much capacity is daunting. That and there's only so much you can do with ~20 base reads (again, the application problem).

I agree with MJC that it's way to early to rule out PacBio.

[Elcannibal]

Officially bankrupt and dead. RIP.

[krobison]

Originally Posted by Elcannibal

Officially bankrupt and dead. RIP.

Note that it is a Chapter 11 bankruptcy & they claim to be debtors-in-possession, so a total liquidation is not (yet) in the works.

[flxlex]

http://www.genengnews.com/gen-news-h...r-11/81247649/