Slashdot Mirror
Celera Maps Entire Fruit Fly Genome
cjoh345 wrote: "Celera Genomics has just sequenced all the genes in the fruit fly. Apparently the scientists involved are amazed at the genes that we share with this dorm-room annoyance. This discovery also validates Celera's "shotgun approach" to mapping out this stuff. And yes, the genome is available free of charge via
Genbank. Good form, Celera!" What would Mendel have thought of this? How about Watson and Crick? This makes me want to break out my copy of The Double Helix .
Hey, that looks like Jon Katz.
Somebody said that if you compare the Genomic mining with the goldrush from the 1850s, Celera is the company selling the maps. Their sister/mother company, PE Biosystems (where I work) is selling the picks and shovels (the sequencers, PCR tools, chemcials & consumables, etc).
:-)
Having mapped the Human Genome will be the platform for about 100 years worth of research, even taken into account the accellerating pace of such research based on technological advances.
During this time, more and more products will come to market, some of which are human drugs, some of which are tailored treatments (e.g. body tissue), some of which are more informational in nature. For this reason, it is very important that politicians such as Clinton and Blair concentrate on these bigger issues at stake - get debates started on what the boundraries are and what we want to do with this information - rather than trying to score populist points by portraying incorrect/ignorant viewpoints about the private sector's intentions in this race. (I guess, they can't make themselves saviors without enemies, but still...)
Anyway - glad the bottle is out of the genie.
For example, this weeks issue of Nature has an article about constructing a model of Parkinsons disease in the fruitfly. This is probably a great tool to test drug candidates on. Fruitflies are somewhat cheaper and more convenient to experiment on than human beings. ;-)
Lars
__
Reality or nothing.
I hereby declare my genetic resources to be open source, and will cheerfully assist any human female that wishes to make use of them.
Oh... sorry, dear, I didn't know you were reading over my shoulder.
I hereby rescind my previous decision re: my genetic resources...
--
--
Don't like it? Respond with words, not karma.
I noticed something a little odd in the timothy's post - he says
What would Mendel have thought of
this? How about Watson and Crick?
This seems to imply that Watson and Crick are as dead as Mendel, which is just not true. Watson is the president of Cold Spring Harbor research institute. His homepage is here. Crick is also alive and doing some rather interesting research in the neurology of consciousness at the Salk Institute. his homepage is here. I met Dr. Crick while working at the Salk, and he's a really nice guy. I've met a number of famous people in my time, but he really awed me. Its hard to talk to him without thinking about the massive influence he has had on modern biology.
If I had to guess, I'd say that they are both as amazed by modern biology as the rest of us. Who could have guessed we'd be this far so soon? Biology is amazing. Computers are amazing. That's why I do both.
ted
Well you know what they say...
"Five a day for better health"
Your Brain + EEG + LEGO Robots = Brainstorms
I've got nits to pick with the rest of what you said, but I'd like to focus on your last remark:
By this, I'm assuming that you're putting the human genetic condition up on a pedestal. Are you really all that confident that things have developed in the best way possible?
Forget chromosomal, mitochondrial, or multifactorial genetic disorders -- serious single-gene disorders alone are estimated to have an incidence of about 1 in every 300 births. Not exactly something to write home about, there.
Cheers,
ZicoKnows@hotmail.com
Correct in some parts. However, what most people don't seem to look at is that organisms are not static. You don't derive the output of a protein solely from the perceived input of a gene sequence. There are other considerations.
I think you hit the nail on the head with this post. I actually thought that the original poster's analogy while true in some senses was slightly confusing. By talking about the genome as being like a physical object solely it ignores the fact that it is, as your post brings out, like a program that has physical parts. This way of talking about it:
It might be clearer to use the analogy of two groups of Victorians being presented with a wonderful little nanobot composed of intricate parts beyond our wildest dreams. The nanobot has been produced by several aeons of amateurish Basic programmers and kindergarten Lego-engineers who have piled fudge upon fudge to improve the thing in response to changing design specs from their PHBs. It is written in a proprietary binary-block format.
One group of scientists decides to study the program that each part of the nanobot carries around, the other group decides to study how the parts fit together and what happens when you take parts out and put them in. Now, the proprietary file format has been decoded and we can look at it as ASCII text. Yee-haww! Oh wait! this is terrible code: goto's , overloaded operators (not that that's always bad, I just hate it personally), no or little documentation. We're going to be a long time figuring out this program.
Calling it the "rosetta stone" similarly is a confusing way of putting it - the rosetta stone had the same text in different languages. This is different texts in the same language. The "words" may be the same in the different "essays", may even have similar patterns of usage in "sentences", but the meaning is damn different of each essay.
Don't get me wrong, I think this is an essential first step (as long as they finish the hard bits...90% only?) but there's a lot of work yet. I'd also like to draw people's attention to the skepticism of some scientists about the assumptions of determinism that are built into this work - you bring this out very nicely in your discussion of the 3D string folding problem of synthesized peptides.[ References:
Not in Our Genes:Steven Rose and Leon Kamin Biology as Ideology:The Doctrine of DNA Richard Lewontin]
If this is the same company that I'm thinking of from a couple years back, they developed this shotgun approach, and were going to finish the fruit fly genome as a test run (it was supposed to be finished by 2000). If they are still following their initial plan, they will now start up on the human genome. They had originally said that after they finished the fruit fly, they'd start the human genome from scratch and still finish it before everyone else. I think they said they'd have it done by 2003. But this is what they were saying probably 2 years ago, so they may have changed everything by now.
Mike
If you create a human, you don't have to make
the source public, unless you DISTRIBUTE it.
Either selling it, renting it, or giving it away.
This probably means we need the source for
hookers, slaves, or married couples.
Perhaps also media-prostitutes, consultants,
and adopted children.
But you don't have to give away YOUR copy, just
let other people clone you.
I can hear all the "anonymous cowards" screaming:
"I want my copy of Natalie Portman, and hot grits down my pants".
Doesn't this reduce the value? I would have thought that the most useful use of a gene map would be to be able to tell which parts of an individual vary, and cause us to be different, which parts are always constant, and those parts which vary only in those who are carriers for gene defects.
If you only map one individual, and that individual has the gene which causes blue eyes, or sickle cell anemia or breast cancer, then that is going to be considered 'normal'. If you map 10 invididuals, and find that 9 of them have the same code and 1 varies, then the one which is found in the majority will be considered 'normal'.
I'm all for mapping the genome, but there are things you have to remeber when you start "tinkering" with it.
:) Excuse me, I have to run off to class now.
For example, did you know there's a whale gene in your tomato? There is, and the reason's it's there is to prevent it from freezing (or at least making it harder to freeze).
Good things: production goes up, losses go down. Bad things: weeds that don't freeze and die in the winter.
This is just a simple example, but some of our crops are genetically enhanced for pesticides. Now what happens when our crops undergo natural pollination and cross breed with weeds?
I read an article about this a few months ago, and I wish I could point out the exact name, or give some links on topics like this, but it was a gardening magazine, and has since disappeared.
This is the reason why Europe boycotts some of our crops.
And for those that are didn't know, they use modified viruses to implement the genes
I love deadlines. I like the "whoosh" sound they make as they fly by. -- Douglas Adams
Someone else mentioned that there was a 3 month waiting period by the company before release.
Does anyone know if the actual genome is publicly available somewhere? I was thinking of republishing it via Project Gutenberg (we've done the draft human genomes that have been released already).
It also needs more work--it's not exactly the entire genome yet.
At this time, ~92% of the genome is in contigs larger than 30kb, and ~78% in contigs greater than 100kb; most gaps are small (3kb or less) and due to genomic repeats, such as transposons.
If the entire sequence is 120Mb, that's a whole lot of pieces they still have to put together. Right now it's already probably good enough to do analysis of protein structure and a lot of the aspects of gene regulation, since they've gotten all the euchromatin, although we already know most of this stuff from the work of developmental biologists.
While the genes they found are interesting and it's great that they actually have sequence confirmation on them, we've probably suspected they were there for years now. Among the more interesting finds was the p53 analog, which controls apoptosis. and whose loss of function in humans is implicated in cancer. But this type of gene is expected to be found in all species that undergo regulative development. The SOD1 analog discovery was also interesting--in humans, this is one of the suspects for the cause of ALS. Its function is to clean up superoxides (perhaps lending credence to the theory that anti-oxidants will let you live longer), and so it wouldn't be that surprising to find something like it in all organisms that do aerobic respiration.
On another note, the progress with sequencing the Drosophila genome may not translate over to the human genome project, because working with Drosophila DNA has a huge advantage: genes are often present in multiple copies within a cell, due to polytene chromosomes. Instead of just having 1 set of double helices per chromosome, you could have 1024 sets or more.
Finally, I just wanted to quibble over the usage of the terms "map" and "sequence." Though they are similar, they are definitely not the same thing. The Drosophila genome could be said to have been mapped thirty or forty years ago, when Morgan did his work. (The units used to describe map distances are named after him) Mapping is the ordering of the genes--determining what chromosomes they're on and how far apart they are from one another. This is often sufficient to determine a lot of things, like inheritance patterns and rates of mutations. Sequencing is just getting the base pairs. A lot of the work will be to get the map and the eequence to match.
Top Twelve Rejected Slogans at Celera:
12. Maximizing current shareholder value for the good of humanity.
11. Have you seen our famous "Geewhiznomics!" roadshow? It's Not Real Science, but an Incredible Simulation.
10. Proving Adam Smith right, one basepair at a time.
9. Now available for licensing: Professionally packaged, Fully Annotated Roman Alphabet Plus(TM).
8. No, Uma Thurman and Ethan Hawke don't actually work here.
7. Celera. Not At All Like Microsoft.
6. "We will bury you, decadent academicians."
5. Just your friendly neighborhood giant corporation profiting from publicly funded reasearch.
4. Have you seen our new island fortress?
3. Proud underwriters of the Aldous Huxley Chair of Biology.
2. Celera. Not At All Like Microsoft.
And the top rejected slogan for Celera....
1. Annotators wanted starting $5.50/hr (nights).
Never throw away good code that works!
Ciao
----
FB
"Analysis of the Drosophila genome revealed 90 previously unknown P450 genes, including those that affect the metabolism of drugs such as beta blockers for heart disease, antidepressants, antipsychotics, and codeine-a pain killer and cough suppressant. (No one knows whether a fruit fly can get a cough.)
Patient to Doctor:"Doc that gene treatment you gave me for my depression really worked well. But you know ever since I've had this strange urge to walk through s#@t."
"Open code, in other words, can be a check on state power." -Lawrence Lessig
Kudos to Celera for releasing the fruit fly genome to the public.
But this brings to the public eye yet again the concept of patenting gene sequences. Yeah, software patents are a concern to me, but patenting the code that makes me what I am really scares me.
I think what is needed is an 'HGPL' -- a Human General Public License, inspired by the GPL.
This is wonderful, and essential. To use an analogy, this is like a Victorian scientist, after years of studying a 1999 notebook computer, managing to deduce how transistors and the wires that connect them work. "
Correct in some parts. However, what most people don't seem to look at is that organisms are not static. You don't derive the output of a protein solely from the perceived input of a gene sequence. There are other considerations.
Even assuming a wonderfully simplistic gene (no introns, single well-defined promoter sequence, well-defined "end" sequence) - all you receive from this is (essentially) a one-dimensional string of peptides. The problem is that the string doesn't stay "one-dimensional" for long. As sections of the string exit from the ribosome, they begin folding into two- and three-dimensional structures, due to environmental factors (an example being how pH affects binding affinity between protein domains). While protein modelling has advanced enormously, it is still wildly insufficient to accurately predict the outcome of domain/finger interaction necessary to generate any but the most simplistic protein structures.
In a nutshell, seeing the peptide sequence for a gene gives you about as good of an idea what the final protein will look like as seeing one bit-representation memory map of a self-modifying program will allow you to determine what the final outcome of the program will be (not to mention not knowing which *parts* of the memory map are actually part of the program and which are parts of the environment ... )
So, yes, there's still *tons* of stuff left to do. But the Victorians will fully understand their computer a heck of a lot sooner than we will fully understand our genome and how such a self-repairing, self-modifying system interacts with a wildly dynamic environment like a cell under stress. :-)
TTYL,
Scott Ferguson
Sanity - coveted most by those who need it least.
Yeah, so they released the Fruitfly source code. Just what we need--soon the OSS community will stomp all the remaining bugs in the bug, and we'll have a stable, reliable, and high performance pest.
On the other hand, I'll bet you a thousand dollars that they never release the source to a really valuable product, like Christy Turlington. Greedy corporate bastards.
Lars
__
Reality or nothing.
I've spent the better part of the last year massaging the drosophila genome with perl scripts. I am intimately familiar with it, and I can tell you that it is NOT complete. While it is 90% sequenced, it is still in a bazillion tiny pieces. This makes it difficult to get complete sequence for many genes, as they may start in one contiguous piece and end in another. From personal experience, I would guess that about 15% of the genes are split like this.
And even if it were complete, this still wouldn't tell us where all the genes are. Its very difficult to string coding regions together into a complete gene, when there may be large introns that confuse the matter. The state of the art still only identifies about 70% of genes correctly, even given complete sequence.
And even if we did know all the genes, we still wouldn't know how they interact. We can make guesses based on previous experimets, but the majority of the genes in a given genome are experimentally uncharacterized. Current attempts at molecular simulations can't even predict how these proteins will fold, let alone with what other proteins they interact, or what they do.
There is still a lot of work to be done - getting the genomic sequence is only the beginning. A significant start, but only the first step. Could you reverse engineer the entire linux kernel if you were only given the binary? Probably not. I assure you that deducing the operation of an organsim given the raw DNA code is much more difficult. The complete sequence of the E.coli (an intestinal bacteria) has been available for a couple of years, and we haven't even begun to understand it. E.coli is single celled, and the genome is only 4.5Mb (thats megabases). Drosophila is very complicated, and the genome is about 120Mb. Don't look for anyone to 'solve the fly' any time soon.
Its a great time to be in bioinformatics - tons and tons of data that noone understands. If we did understand it all, I'd be out of a job. I'm not worried.
ted
human genome (more than 200trillion base pairs)
A bit over 3,000 million, actually.
To compare, the government-funded Human Genome Project has so far spent over 10 years on the same job.
There's been a lot of mapping done, that Celera is actually using at one remove, because they're picking up our data for their assembly. Celera's job would have been a lot harder if our work hadn't been available to them.
Celera is doing a 4x oversampling on the human genome, unlike the HGP, which does 10x oversampling. This is possible because Celera is sequencing DNA from one single individual (most likely Craig Venter?), thus avoiding the uncertainty of wheter differences are due to sequencing artifacts or personal variations.
The extra depth is more to do with accuracy. Incidentally, does anyone know when Celera ditched their aim to collect lots of variation data? That was going to be their great contribution to human knowledge, and their main selling point. When did they change their minds?
Plus, Celera is using our (PE Biosystems) 3700 DNA Analyzer (fully automated, unattended operation 24 hours per day), whereas the Human Genome Project mostly use our older 377 DNA Sequencer, which requires manual reloading of samples after each run (every 2-3 hours).
As I've pointed out, here at the Sanger we've got more than 100 3700s : other institutions have gone for MegaBace machines instead.
I guess the next logical step would be to mass produce giant mutant fruit flies the size of cows and then harvest them for food?
From their press release....they have a map, they don't have the sequence yet. A map is a guide with landmarks as to where major chunks of sequence fall within the genome. Celera has pioneered the approach of shotgun cloning. They randomly capture chunks of the genome and sequence them. If they do this enough times (10-50X genome size) they will ultimately have the entire genome after some sophisticated algorithyms sort the data and place it onto the map. They probably have most of the genome and have a few difficult bits to figure out (some sequences are harder then others to get).
Some searching reveals the NCBI press release. Looks like they have most of the sequence together. I'll bet berkeley provided the map which is allowing Celera to put their info together.
The big question is: Has Celera already filed patents on every ORF it has found; Will the patent office grant the patents; Will Celera get patents on the human homologs of these genes (they have identified most of the homologous sequences from EST's in the human genome project).
And you thought software patenting was fusked.
no sig.
You are absolutely correct! The next misson (after sequencing the genome is to figure out what all of those genes(proteins) are doing. So, onto the human structome project (structures of every protein that has been inferred through sequence) and the functome (functions of all genes that have been inferred from sequece data). I think 50 years is a resonable estimate. Things may accellerate a little once physicists (and coders) get involved and start creating models of signaling networks within cells. Models that predict real cellular responses to the additional expression of genes X & Y will be the one that begin to understand how cells operate.
Just think how different coding will be 50 years from now.
no sig.
Fortunately, things aren't quite as bleak as you portray them. Many, many proteins do in fact have a single, clearly defined primary function, either by themselves or as part of a larger complex. Those proteins can have their function inferred either by watching them catalyze reactions, deleting them and seeing how it affects celluar function, or comparing them to similar proteins from the same or other species.
More promisingly, new techniques of functional genomics and proteomics are being developed to analyze protein function by looking at more subtle factors. To find the function of a protein of unknown function, you can find out what other proteins it interacts with and infer what role it plays. You can also grow cells or organisms under different conditions and look for changes in levels of gene or protein expression to determine what proteins are associated with specific metabolic or other life states. Some very interesting work is also being done by determining the 3D structure of proteins (either by analysis or simulation) and predicting function based on structure.
The tools for the next big thing are out there. It's just a matter of going through the long grind of applying them. It's going to be a very long road, probably much longer than the process of sequencing the genome, but finding out (to a rough and ready approximation, at least) what every protein does is an accomplishable goal.
There's no point in questioning authority if you aren't going to listen to the answers.
Actually the sequencing part was completed earlier this year (giving tremendous subsequent rise to Celera stock). What they now did was the mapping - i.e. piecing the small fragments together.
:-)
Moreover, Celera have also completed more than 90% of the human genome within the last year or so. Once complete, this is an indication that the time before the human genome (more than 200 trillion base pairs) are mapped will be shorter than originally anticipated. To compare, the government-funded Human Genome Project has so far spent over 10 years on the same job.
Celera is doing a 4x oversampling on the human genome, unlike the HGP, which does 10x oversampling. This is possible because Celera is sequencing DNA from one single individual (most likely Craig Venter?), thus avoiding the uncertainty of wheter differences are due to sequencing artifacts or personal variations.
Plus, Celera is using our (PE Biosystems) 3700 DNA Analyzer (fully automated, unattended operation 24 hours per day), whereas the Human Genome Project mostly use our older 377 DNA Sequencer, which requires manual reloading of samples after each run (every 2-3 hours).
As originally stated when Celera was created two years ago, the data is going to be publicly available - a point that has gotten lost among very opinionated but not so informed readers of Slashdot. There will be a 3-month lag period, to ensure accuracy of the data, and to see if there is any information that could be used for patentable drugs & applications.
(Mostly, Celera's business model is based on providing the tools that will give access to this database).
And I have stock options!
-tor
Here is a much more detailed link of the story from Celera's site, talking about the similarities between our genes and the fruit fly's. (I've got a dollar that says their computers are all Celerons, ha ha ho ho.)
What's your damage, Heather?
Celera hasn't even by their own definition sequenced the entire genome of Drosophila. What they have done is sequenced most of, or all of the euchromatic region. The highly repetitive heterochromatic DNA that is clustered around the centromeric regions and makes up an estimated 30% of Drosophila genome is not sequenced. There may even be some B-heterochromatic regions which are also unclonable which is a serious problem in trying to sequence highly repetive DNA. While these regions don't have the glamour of the gene-rich euchromatin (it is often referred to as junk DNA for that reason) they can effect everything from gene expression to chromosome pairing.
And to the few posts I have read which think that this is some sort of private enterprise success story versus a slow blundering government, it isn't. It is a classic example of business going after the highly profitable bits and leaving the taxpayer to fund the basic research. The same basic research which incidentally made it all possible in the first place.
This is a wonderful accomplishment - now we can get started on the main problem.
Having a complete map of a creature's DNA tells us, in principle, all of the proteins that it can synthesize throughout its lifetime. This gives us the building blocks that the creature uses to build things, and the chemical signals that it uses to direct internal operations.
This is wonderful, and essential. To use an analogy, this is like a Victorian scientist, after years of studying a 1999 notebook computer, managing to deduce how transistors and the wires that connect them work.
He still needs to deduce a lot about capacitance, resistance, and inductance to tell how signals will propagate and influence each other, and needs to build up from scratch all of the disciplines involved in integrated circuit design before he can understand how it works, but it's a start.
Similarly, we can now move on to the next step in understanding biological creatures - trying to figure out what all of the proteins do, and how the systems built from them operate and interact with other such systems.
This would not be an easy task under the best of circumstances. It's made worse by the fact that evolution puts little value on modularity - the systems will interact with each other to such a degree that it will be difficult to even define individual systems within the chaos that is an evolved being.
I wish them luck. They have opened the door, and made available for study the vast landscape of interacting systems that we'll have to understand to truly understand how living creatures work.
-----BEGIN FRUITFLY-----
ATCGTAGCTAGCTGACTATCGTAGCTAGCTAGCGTATCGTAGCGATCG
GCATCGTAGCTAGCTAGCTACGTAGCTAGCTAGCGATCGTACGATCGA
CGTAGCATCGTAGCTAGCTACGTACGATCGATCGATCGTAGCTAGCTA
GACTAGCTAGCTAGCTAGCTAGCTACGTAGCGCGATCGATTCGATCGC
AGCTTGACTGATCGGATCGTGCTACGGACTGTACGATCGTACGATCGC
------END FRUITFLY------
GATCGATCGATGCTAGCTACGATCTGATCGATCGATCGTAGCTAGCTA