Slashdot Mirror
Open Source Biology And Knowledge Distribution
n7lyg writes: "IEEE Spectrum has an opinion piece this month on Open-source Biology and Its Impact on Industry. The article speculates about advancements in biology and biological-based manufacturing and how it is likely impossible for anyone to keep control of the Intellectual Property around genetic engineering. If software was as easy to produce and prevalent as DNA, then it would be similarly impossible to control software as IP."
Ok, I understand the tools to make tools arguement and the pace of change and all that. But in some ways that makes my point even more appropriate. Despite that effect we still have no strong AI and are still looking at estimates of (at least) yet another decade before we do. You would be correct if you detect in my tone a bit of skeptisism on this yet latest estimate.
Genetics are at least as complicated as the brain and we have essentially no experience with them. That approximation may offend some scientists in the field, and I don't mean for that, but look at were we are. What are there, a couple tests for genetic diseases, a kind of corn that keeps leaking into the mainstream supply and an identity test that couldn't convict OJ?
We had a similar level of cognitive understanding a couple hundred years ago. Certainly Turing understood at least that much more recently. Yet we still stumble.
Couple that with the fact that it isn't 'simply' genetics, but also the surrounding cellular chemistry that makes the genes work that needs to be studied.
There is an awful lot to understand and we don't even come close to having the capability to do so yet.
So, I'm just saying that it's easy to get hyped over this stuff, but be a little cautious about when you promise your deliverables. Ask Minsky why if you still don't believe me.
--
Poliglut
They were, obviously, wrong. Today we can't even do speech recognition with sufficient reliability to use it every day, much less get in a car, tell it were we want to go and then read the important news of the day (pre-selected by a Chevy Agent) on the drive.
So too will genetics be. We're at the beginning of an interesting period in research. But for the genetic possiblities to be realized will require something like the kind of AI imagined in the 60's to be available and well understood. There is just that much data available to process.
So, while I'm as excited as the next guy at the possibilities, I'll consider myself lucky indeed if I live to see them realized.
--
Poliglut
I find this article just a little too forward thinking and optimistic. I'm a experimental medicine PhD student, and to be honest, I'm pretty sick of hearing all the hype about how knowing the genomic sequence is going to sling shot us into understanding all biological interactions in the next 50 years. Gimmie a break! I really think where we are at this point fits a historical anology:
Discovery of the Rosetta stone in 1799 ( http://www.rosetta.com/RosettaStone.html ) allowed historians/archeoligists to finally understand Egyptian hieroglyphs. Did these historains then understand Egyptian customs, culture and building techniques within the next 50 years? No WAY! Were untrained indivduals, so called hobyists all performing cultural analysis of ancient Egypt? I doubt it. It did, however, launch an era of increased understanding for this lost culture.
We are accelerating in our aquisition of knowledge regarding biological systems, and the genome projects currently underway will no doubt facilitate further acceleration. But knowing the genitic code doesn't mean that everybody is going to be creating transgenic organisms in their garages in 50 years.
Now that we have the genetic code (and remember, it turns out we have FAR fewer genes than we anticipated, 30,000 vs the expected 100,000) the next great leap forward in biology is going to be study of protein protein interactions. How fast will we progress in this regard? To give everyone a yardstick, the gene responsible for cystic fibrosis was identified (without the help of the human genome project) 10 years ago. Are we any closer to a cure for cystic fibrosis? Doesn't look like it (despite the efforts by fantastic researchers in the field). That is one gene, that happens to cause one disease. Its becoming an accepted fact (or at least a paradim) that for many diseases, or biological processes (in the case of biological manufacturing) several genes and therfore proteins interact to produce other products.
A LOT of further study is required in order to understand these interactions, its gonna take a little over 50 years (mabey 100+) for us to understand biology in the manner this author describes. Right now, the following steps need to occur:
1) Linkage of ORFs within the DNA (open reading frames) or full genes (expressed mRNA transcripts) with the proteins they code for. This still needs to be doen for some genes, as the human genome project (as opossed to Celera's efforts-which are "closed source") releyed on sequencing all the DNA as opposed to expressed genes.
2) Study of gene expression and regulation. Much of the DNA sequenced in the human genome project is called "Junk" DNA. In fact, these sequences contain mostifs, called promoter (and in some case suppressor)sequences which regulate the timing of gene expression- in many cases things are produced, or not produced, singnals dampened or aplified based on the timing of gene expression. Much of this work will be performed utilizing "gene chip" technology, for high throughput screening.
3) Study of Protein protein interactions. Much of this work is performed using either immunoprecipitation (mixing together cellular proteins, pulling one your interested in out of solution using immunoprecipitation and seeing what other proteins are stuck to it) and the yeast two-hybrid technique (a very kickass technique for high throughput screening of genomic libraries to identify interacting proteins).
4) Confirmation of function utilizing genetic knockouts, inhibitors or other transgenic model systems.
A lab could easily spend 10 years on performing all of this work on one or two genes of interest, and it would be unlikely they would figure out all the protein-protein interactions. This represents the work of 5 scientists ( a conservative estimate) X 10 years= 50 years of "people" power work per gene. Multiply this by 30,000 (number of human genes identified) =1.5 million years of work. Muliply this by the number of organims you expect to "know" the biology for,(50 is a low estimate) =75 million years of work. Divide this by the number of scientists, and you have the time needed to "understand" all these genes. From my very rough estimation, we'd only need 1.5 million (Very well funded) scientist to complete this work in the next 50 years. Ya right. Don't be expecting a biological version of the replicator for another 100 years at least.
"If we knew what we were doing, it wouldn't be called research, would it?" Albert Einstein
"If we knew what we were doing, it wouldn't be called research, now would it?' -Albert Einstein-
Of course, any statement on what the world is going to be like in 50 years is purest speculation. This story, however, is even pushing that characterization a bit, as it contains several exceptionally questionable assumptions.
Foremost is the idea that people will soon be effortlessly tinkering with genetic modification in their garages in the near future. The evidence of the availablity of home-made DNA sequencing techniques like gel electrophoresis is cited. This is a very dated and limited technology, available as a kit to any high-school biology teacher well over a decade ago. It is an incredible far cry from the kind of DNA analysis on a chip that is in current play in the research sets. The question is, are the powerful corporations who are staking their existence on capitalizing on genetic information going to quietly watch this kind of technology get domesticated, or are they going to fiercely lobby to keep access restricted to the government and corporate level, most likely argument that the dangers of making genetic modification avvailable at the consumer level are simply too great because of the potential for use in terrorism (i.e. biological warfare).
If that's a bit too conspiracy for you, consider the fact that the corporations are far ahead of the curve in the development and posession of these patented technologies. What are they doing with this power? Why, they're deducing functional bits of DNA... and patenting them as fast as they possibly can, usually with some pretty spurious uses attached to justify what is essentially gene-squatting. An Open Source model in genetics is well into the process destroyed by an ignorant, corrupt and complacent government and the fast actions of dozens of greedy corporate giants and hungry start-ups.
This is to say nothing of the incredibly optimistic timeline and assumptions this article makes about how effective "hacking" biology for useful industrial processes is going to be.
My favorite part is the examples given of garage technologies- automobiles and computers. Ah yes... It sure is great living in this capitalistic paradise where cars and computers are being innovated by thousands of inventors using a non-proprietary, open intellectual property model. Think of the horror of living in a world where a handful of corporate monoliths produced the vast majority of vehicles and microprocessors, enforcing proprietary ownership of their technologies through international patent laws!
It Is the Nature of Information to Transgress Artificial Boundaries
Working as I do, a software engineer in a bioinformatics research institute, I can tell you that even at the frontier of research people have no clue as to how some things work, let alone how to change them.
As a computer vision expert, David Marr, said once, "the tools we have for describing nature are very weak." Notice he said describe, not understand or model! For the while, we'd be happy to find that some genes can be used as markers of activity, without knowing the mechanism of that activity at all. IMHO, this fifty-year-forward vision of biotech has some orders of magnitude lacking...
I don't mean it's impossible, it's just the relations involved are much less straightforward than in software. And remember we didn't design it. To find the structure of our own spaghetti code is hard enough, to understand and improve nature's design over millions of years will be an immense task. But the task is ours to tackle!
make DNAImage, anybody?
Linux *is* user friendly. It's not idiot-friendly or fool-friendly!