[krobison]

To chime in quickly: I too would be very pessimistic about the business of bioinformatics (which is different than the profession of bioinformatics!). Many of the same arguments Joann is making were made in the go-go years of the first genome wave, and lots of companies were founded either directly around bioinformatics or selling databases where specialized bioinformatics added big value. Many of those companies are either completely evaporated (e.g. Pangea), shifted out of bioinformatics (e.g. Compugen), or merged into somewhere else -- and those somewhere elses often seem to be hanging on by their fingernails. There are a few successes, but very few.

One problem was identified by Anthony -- there is just too much risk that open source projects will eat your lunch. It's challenging to compete with free. Even if you can build & sustain a superior product, a lot of potential customers are going to go with free, figuring that the delta between the two is better spent on something else. There's also the problems that the goalposts not only drift forward, but sometimes lurch sideways. If you look at Microsoft Office from a decade ago, it's pretty recognizable as the ancestor of today's Office. If you look at the big bioinformatics problems of a decade ago, they have similar themes (genome assembly, RNA->genome alignment), but the algorithmic landscape is totally different due to the sequencing technology shift. There's a lot less infrastructure you can recycle.

One last comment: I'm surprised by the remark that there hasn't been much activity in subsetting the genome for 2nd gen sequencing. There's been at least a dozen papers in the last few years with several different technologies, and most of those papers are probably in the last 6 months. By my count there are five vendors offering technology in that space. While the push for even more productive sequencer runs will continue to drive down the cost of doing a genome, these selection methods will still have utility for processing large numbers of samples.

[apfejes]

Hi krobison,

Thanks for bringing up some great examples and joining into the conversation. I just wanted to add one comment to your last point about sub-setting the genome.

I think there are several reasons why we aren't seeing more of it: The first is that there are already decent capture technologies that are relatively agile that they can be applied in various ways (custom arrays are probably one of the most popular of the bunch) while various other "less biased" methods are available (RNA-Seq of various types) that do a moderately reasonable job.

However, I think the real issue is just that the technology is still too much of a moving target - if any company stakes their position on a capture technology for Illumina/454/SOLiD platforms, they're likely to be blown away in the next year or so when Pacific Biosciences or Complete Genomics or whatever is next emerges. I suspect that the really innovative genome sub-setting technologies will only pop up once we have a stable platform available upon which other companies/labs can build upon.

Anyhow, just my $0.02, since this post is entirely speculative. (=

[Joann]

Hi to all,

Keep in mind that this is an interdisciplinary type discussion. What I am hoping to point out is that the key technology of next gen sequencing is entirely a creature of computational methods, rather than a wet lab approach. My observations on fractionation (you say subsetting) are based on the decades of physical subcellular fractionation techniques that might have supported nuclear (of the cell nucleus) separation in a reproducable manner and that could be well validated by sequencing. However whole cell sequencing would be able to get at the DNA dynamics with much greater reproducability. The best example I can think of as fractionation is ChIPSeq.
And on the commercial side, with patent conventions as they are, straight sequencing of organisms for utility genes probably can support business models. The key difference between now and then is that algorithms rule and the physical/chemical sequencing procedures can't get you there without them.

[rranauro]

Very intriguing, allow me to jump into this thread. (Apologies for the long post, this is my first post)

First, to answer smice' question of market potential the premise has to be that over time, deep sequencing will make inroads in to existing research markets for gene expression, variant discovery, gene regulation, transcript discovery, and so on. Who will consume that information? Excluding the big research labs, you can count 300-500 medical research labs, 300 large commercial pharma/bio/agro enterprises, and 1500 small biotechs, and 4000 academic core labs world wide.

Investors want to know the AVAILABLE market as the sum commercial potential and the SERVED market as the sum commercial existing sales to the market. Suffice it to say that research bioinformatics is largely an in-house endeavor comprised mainly of the fully loaded cost of the human resources providing bioinformatics answers to research questions.

As for the business of bioinformatics, I would tend to agree with the proposition that bioinformatics as a TECHNOLOGY is a tough business to be in. However, products and services that aleviate drudge labor and allow these energies to be re-deployed on more value creation activities should have a market. Competition with open source just raises the bar on innovation comprising not just technology but business model and strategy. I would argue that the early innovators in Bioinformatics did not sustainably do that in all three dimensions and therefore are not positioned to capitalize on this emerging market opportunity.

[steven]

Hi,
Thank you all for this interesting thread. I think that this short survey could help. It is about the needs of the community in terms of bioinformatics tools for NGS data analysis.
cheers,
s.

[apfejes]

Hi Steven,

If you're going to post a link to a survey, could you at least tell us who will be collecting the data? I'd like to know if that's a corporate survey or just for your own entertainment.

Thanks,

Anthony

[ECO]

Hi Steven,

If you're going to post a link to a survey, could you at least tell us who will be collecting the data? I'd like to know if that's a corporate survey or just for your own entertainment.

Thanks,

Anthony

Heh, I just asked him the same thing in his other thread about the survey. Methinks there's more than meets the eye here.

[steven]

Hi,
Sorry for the lack of context, I did not want to post a long text again. I have to admit that entertainment is not the only interest I have in next generation sequencing. There's scientific curiosity too (as for everyone here). More information is available in my previous post.
Maybe it would be better like this:
DISCLAIMER 1: This survey may be used for commercial purpose (pretty much like any other information in this forum I think).
DISCLAIMER 2: As I am not planning to prevent people from academic institutes to access the results, it may be used for non commercial purpose too.
DISCLAIMER 3: It can only be used if people do fill it!
So please have a look, it is very short, and hopefully entertaining! Listen to your scientific curiosity...
Anyway, thanks for any feedback.
s.

[apfejes]

Hi steven,

I think, in general, people just object to collecting information when we don't know who will end up using it and for what purpose.

The fact that you're clearly doing this and intend to use the information to help you with your current position at an undisclosed company makes it that much less transparent. Since most of us are academics or work in an academic-related environment, we expect people dealing with us to be as transparent about their goals as possible.

I appreciate your reply - but I don't think I'll be taking your survey any time soon, anyhow.

[steven]

I think, in general, people just object to collecting information when we don't know who will end up using it and for what purpose.

The fact that you're clearly doing this and intend to use the information to help you with your current position at an undisclosed company makes it that much less transparent. Since most of us are academics or work in an academic-related environment, we expect people dealing with us to be as transparent about their goals as possible.

Hi Anthony,
Please note that from my first post I have been more transparent about my goals than probably 95% of the people here. That is precisely because I no longer work (exclusively) in the public domain that I felt I had to precise my intentions. It would have been so much easier for me to say something like "Hey, we are grad students working on a practical summer project with a biotech, please help!" (shorter and more efficient -not mentioning that on top of that, we sometimes happen to have students doing exactly this kind of marketing training, so that could perfectly happen). Now if you really want to know my name and my company I can tell you offline, as I said I am not here to sell/advertise. Plus it would allow you to check my -modest- publication record, mainly dedicated to academic research.
Anyway, I totally understand your reaction... I would probably have the same! Except that I would have given a try to the survey of course
My point is that I believe/hope there is a need for good software that could be used by biologists with no computational skills. And I also believe/hope there is still a gap between the available commercial solutions and the current needs. And I also believe/hope that there is another market for a bioinfo software company than to only rely on selling huge and expensive products to a few customers like big pharmas, as it was earlier said in this thread I think (maybe I am just being naive).
So, I made this survey to improve all the above (including naivety).
To conclude about my personal motivation, as a researcher, my goal is simple: help increase human knowledge... even if "so does the shore of our ignorance", to quote Wheeler's paraphrase of your Aristotle signature.
Sorry for the length of my posts.
Yours,
s.

[apfejes]

Hi Steven,

I hope you didn't take my response personally, as it wasn't intended to cause offense. However, I would like to rebut a few of your points.

First of all, I think 95% of the people here are clear and honest about their affiliations, and I would (perhaps naively) think that their goals are also relatively clear. We are all in the game of getting answers as quickly as we can - and beating our fellow researchers, but I would like to believe that we are also a community.

If you had lied and told us that you were a grad student, I don't think it would have furthered your cause - your survey is clearly commercial in nature, and most of us here are well versed in sorting out the "wheat from the chaff."

Finally, I think I'd like to point out that I'm also in the business of providing good software to biologists with little or no computational skills - however I'm trying to do so as a community effort, rather than commercial. In the grand scheme of things, however, one might ask as to which of us is really doing it to "help increase human knowledge?" The project that is done as a commercial venture, or the project that's done as a open community effort?

I'll leave your point about there being a market for a bioinformatic tool kit alone, only to point out that I've provided my thoughts on that point on my blog and elsewhere in the forum, if you're inclined to discuss that further.

Anyhow, I think my main point stands, as long as I don't know who it is that will be looking at the results of your survey, I'm disinclined to participate. Other scientists may feel differently, and that is their choice. I'm simply using the forum to engage in discussion - and I appreciate your continuation of the same!

Anthony

[steven]

Anthony,
no offense. I actually find the discussion interesting, and I hope more people do. For once I'll be brief: I try to believe that I am still helping scientific research, although working in a private company. More here.
Regarding the survey, preliminary results are already available here!
cheers,
s.