IT Spending in Life Sciences

dano1992 writes "From Cnet: Computers replace petri dishes in biological labs. "The life sciences field is poised to spend billions on IT due to a need to manage an explosion in biosciences data, and a desire on the part of drug companies to streamline drug development." But the folk who'll catch the best part of the wave are those who can work with clusters, databases and storage on a massive scale."

So...

[Roto-Rooter Man]

Are the drug companies evil this week or not?

Nobel Prizes

[henrygb]

will go to those who ask their databases clever questions. Getting the answers takes skill but is not at the same level.

Not gray goo!

[Phronesis]

Leave it to CNet to revive Bill Joy's spurious fear of gray goo taking over the world. Get over it, Bill. To paraphrase Stephen Jay Gould, It's here, it's bacteria, get used to it!

IT enables, not replaces

[airuck]

"Computers replace petri dishes in biological labs."

That isn't very realistic. Data collection is still the major driving force in life science discovery. Good IT infrastructure enables large screens, but only in conjuction with robotics, microfluidics, sweat, and a lot of disposable plastic, including petri dishes.

Modeling biological systems is a difficult task. As Hiroaki Kitano [cvbig.org] points out, "[in biological systems] large numbers of functionally diverse, and frequently multifunctional, sets of elements interact selectively and nonlinearly to produce coherant rather than complex behaviours". There are still a huge number of elements and relationships to discover.

Yes... and no.

[carldot67]

There are lots of helpful IT tools around that are used in drug discovery. These range from simple sequence alignment algorithms right through to complex ab initio protein folding code - ie quantum mechanics.

While these programs are very helpful, they often contain shortcuts that reduce compute time. More compute crunch essentially means that these shortcuts can be removed and deeper/wider/more accurate analysis results. Which is good.

That said, more raw crunch and capacity brings in other issues such as capacity, I/O, network, concurrence, version control, security, recovery, UPS, climate and so on and so forth. The new iron, in other words, needs looking after. So some of the new hardware is there simply to look after the other hardware, if you get my drift.

Remember, there is more to drug discovery than meets the eye. Living systems are extremely complex. Drugs or hypotheses that look great in silico do not always pass muster in vitro never mind the real world. Moreover, FDA approval still relies on squirting compounds into cells, rats, humans etc. Until the FDA permits in silico proof of efficacy, toxicity, LD50 etc, we will need to maintain "traditional" avenues for experimentation.

It is good stuff, though and I for one welcome the investment in new compute capacity. I am keen however that no-one is seduced by the headlines; a lot of hard work has to be done in the lab to corroborate the evidence uncovered by the computers. In my experience the data is academically interesting but is only the beginning in terms of delivering an effective therapy to the patient.

Not knocking it - merely a reality check!

Open Source molecular biology software

[holy zarquon's singi]

A lot of the molecular biology software is open source. Much of it is unlicenced. Alot of the software available is bereft of software licences. Some has weird licencing conditions, and can be difficult to obtain for non commercial purposes. Many molecular biologists have a gap in their computer literacy. I think that there's definately a market for a knoppix style cd that includes a large suite of molecular biology tools, database software (probably postgresql) and good documentation, so that the poor molecular biologists who've been suffering from using windows all these years can return to a real scientific computing environment. There are many good analysis tools available for unix, but few good presentation tools. A nice genome drawing package that outputs [encapsulated] postscript, as well as documentation explaining what these weird file formats are for would be really good. The combination of a self documenting self configuring linux distribution with built in clustering support would be very well received if the scientists could be eased into changing their computing platforms.

Re:Open Source molecular biology software

[Bowling Moses]

A lot of it is open source; unfortunately a lot of it also epitomizes what's wrong with non-commercial software. You might have just one small group, one lab, or even just one guy who's little baby that program is. For what I've used, the majority of the open source software and freeware runs on unix, not windows, although there is a growing commitment to linux and windows that will become dominant in the next couple of years. Most of it runs. Sort of. If you're willing to fight with it. The only thing worse than the user interface is the manual, if one exists at all (although exceptions exist, such as Deep View Swiss PDB Viewer [expasy.org]). While I'd agree that many biologists have poor computer skills, that's rapidly changing as more and more the problem is data interpretation as opposed to data acquisition. There are many subdisciplins where a good percentage of the biologists also possess some ability for computer programming (or at least script writing) just becuase that's the only way to force the software to work. I'd also agree that a CD package with documentation on a popular set of programs would be great, but that's complicated by the fact that the geneticists don't do biochemistry who don't do cell biology etc. That and getting a bunch of scientists together on something like this is like herding cats.

looking for a job? here' one.

[Anonymous Coward]

I know of a technician opening in this exact area, the pay is low by san jose standards but then so it the cost of living in new mexico: 60-80K$/year for someone with LARGE linux cluster sys admin, and bio-tech experience. Preferably a US citizen who did a pHD level work in molecular dyamics or biological structure modeling and is now burned out on the paper chase and grant writing and now wants a cozy technician job in a scenic part of northern new mexico in a very large government lab. Needs to be able

Managing the mountians of data

[Galahad]

All you need is NuGenesis Technologies' [nugenesis.com] Scientific Data Management System [nugenesis.com], or SDMS for short (don't you love acronyms?) to manage [nugenesis.com] all of your analytical data. It won Best In Show at Bio-ITWorld [bioitworldexpo.com] this spring.

Disclaimer: I work there.