Slashdot Mirror
Creating Artificial Proteins
Spy der Mann writes "By examining how proteins have evolved, UT Southwestern Medical Center researchers have been able to design genes to create artificial proteins.
The researchers have discovered a set of simple "rules" that nature appears to use to design proteins. By feeding these rules into a computer program, they were able to obtain a sequence of artificial genes. These genes were then inserted into laboratory bacteria, producing the artificial proteins as expected."
I've been creating proteins by hand since I was 12!
Norman Cook's Ode to Sl
Hopefully we'll figure out what not to say while learning the grammar, style, and syntax of this new language. It's a bit worrying -- but this really has a lot of potential, I think!
Paleotechnologist and connoisseur of pretty shiny things.
Isn't one of the reasons that creationist use when they attack evolution (actually abiogenesis) is that it would take such a long time to generate functioning proteins through random chance that it would be statistically impossible? If there are simple "rules" to create proteins, maybe that's how nature was able to come up with life so quickly.
Real_men_don't_need_spacebars.
Good lord, what if people had other unique markings that could be tracked... finger/palm prints, DNA, retinas... now THAT would be scary.
Oh wait.
"Hel-l-l-p me-e-e-e-e..."
Seeing bad movies only encourages them. Watch responsibly
Well, we know we've been able to modify DNA to produce insulin from bacteria.
We've got bacteria that crap out metal wires (Can't remember if we discovered them or made them)
Now where's the bacteria that will make substances like xanax or other drugs, so it can make the entire market cheaper and more affordable to those who need it but don't have insurance, and "naturally" at that? (Naturally as in not needing a buttload of power from a processing plant for the drug and wasting energy uselessly)
Still waiting on Serviscope_minor to wake up to fucking reality and realize that Jessica Price isn't going to fuck him.
I think you might be using backwards logic here. TFA states that by examinig 100 proteins they were able to notice some standard common things about the proteins they were looking at. When they made rules around those common things they could make new proteins.
It's like having 100 pieces of example code to look at before trying to create your own, not generating the code from nothing.
s
And the answer to this could be that a lot of rules have been randomly tried out. It turns out that the rule(s) we are seeing/discovering are the ones that lasted - and if they are simple they are probably efficient in some way.
The creationist/ID policy is to avoid facing unknowns by passing the buck onto a designer. In the current example, just because something appears elegant and simple to some person, it does'nt mean that it could not have naturally occured.
Our jobs, as scientists, or in the more general case, as people with a scientific temperement, is to uncover how or why this simple and elegant thing is the way it is - not to say, 'It's too tough, lets pass the buck onto the designer'!
The researchers believe they may have found a set of statistical rules for determining the tertiary ('overall') structure of proteins from the sequence.
(Although the summary reads otherwise, creating a 'new' protein with an arbitrary amino acid sequence isn't new at all though. )
If this pans out, it is of course significant towards the goal of engineering 'new' proteins one day. But there is still a lot to be covered. Even if the relationship between sequence and structure were simple and known (and it isn't, yet), you still have the issue of relating structure to function.
Which isn't known. And of course, even knowing the structure and function of a single protein doesn't mean you know what it's going to do in a complicated environment such as a cell, where there are thousands of things to interact with.
It's a step forward, nonetheless. But if someone thinks this means we're going to be tricking-out living organisms with new custom-engineered proteins anytime soon, you'll be disappointed.
Artificial proteins! YES! One step closer to Artificial steak!
Can this be used for information compression in any way? After all, it was discovered about 20 years ago that simple fractal equations gave shapes very much like ferns. This could give you a shorthand way of compressing the genome of an organism, then making comparisons.
It would also, of course, be interesting if you could use this to work backwards through the genome to a set point, and (hypothetically) bring back the Auroch.
Personally, I want to see how this deals with metal incorporation at the active site, and whether their selection rules work for that as well.
the more accurate the calculations became, the more the concepts tended to vanish into thin air. R. S. Mulliken
How long before gcc supports this new instruciton set? :p
Full text of article, institutional/personal subscription required.
Abstract: Classical studies show that for many proteins, the information required for specifying the tertiary structure is contained in the amino acid sequence. Here, we attempt to define the sequence rules for specifying a protein fold by computationally creating artificial protein sequences using only statistical information encoded in a multiple sequence alignment and no tertiary structure information. Experimental testing of libraries of artificial WW domain sequences shows that a simple statistical energy function capturing coevolution between amino acid residues is necessary and sufficient to specify sequences that fold into native structures. The artificial proteins show thermodynamic stabilities similar to natural WW domains, and structure determination of one artificial protein shows excellent agreement with the WW fold at atomic resolution. The relative simplicity of the information used for creating sequences suggests a marked reduction to the potential complexity of the protein-folding problem.
From this page : a WW domain is the smallest, monomeric, triple-stranded, anti-parallel beta-sheet protein domain that is stable in the absence of disulfide bonds, cofactors or ligands.
Why can't these articles include any meaningful information? They refuse to tell you what they're about.
Earlier research has shown that for a given group of related proteins, or protein family, all family members share common structures and functions.
What would be an example of a "protein family" in this context? Filamentous? Membrane associated? Globins? Antibodies? No idea. "Common structures and functions" could mean several different things.
By examining more than 100 members of one protein family, the UT Southwestern group found that the proteins share a specific pattern of amino acid selection rules that are unique to that family.
This tells us nothing that isn't already known. Of COURSE proteins with related functions share specific patterns of amino acid selection rules or they wouldn't work. WHAT sort of selection rule did this group actually find?
"What we have found is the body of information that is fundamentally ancient within each protein family, and that information is enough to specify the structure of modern-day proteins," Dr. Ranganathan said.
He sounds like he's talking to a little kid.
He and his team tested their newly discovered "rules" gleaned from the evolutionary record by feeding them into a computer program they developed. The program generated sequences of amino acids,
and how did it do this?
which the researchers then "back-translated" to create artificial genes.
i.e. they did a trivial replacement of single amino acid letters with three letter codons in silico, then generated the corresponding DNA sequence.
Once inserted into laboratory bacteria, the genes produced artificial proteins as predicted. "We found that when isolated, our artificial proteins exhibit the same range of structure and function that is exhibited by the starting set of natural proteins," Dr. Ranganathan said. "The real test will be to put them back into a living organism such as yeast or fruit flies and see how they compete with natural proteins in an evolutionary sense."
Translation from stupid-articlese: in vitro the translation products of the artificial DNA folded into shapes similar to wild type proteins. I think.
One can only assume that these guys chose proteins that don't undergo post-translational modification.
Most proteins eventually degrade, if they are not immediately destroyed by the immune system (ie, antigenic). Furthermore, for proteins that don't degrade quickly, how would you detect these proteins? Other than putting radioactive isotopes (try getting on an airplane with that in today's environment!), I don't see how you would detect them other than strapping someone down and getting some blood. I suppose you could always try a gene therapy technique to continually express protein, but gene therapy is still highly experimental and presents its own problems. This sounds way more complicated than just implanting inorganic RFID chips/beacons/whatevers under the skin or in a (cough!) body cavity.
maybe now all us anime/manga fans can rejoice and change the hair color of our kids to sky blue, sea green, or bubblegum pink, or some other outlandish color like purple
Viable Slashdot alternatives: https://pipedot.org/ and http://soylentnews.org/
If that was all they'd done I find it difficult to see how this differs from doing a multiple sequence alignment for a family of proteins, then making a gene for the consensus sequence.
Checking the paper (and related News and Views article) in Nature itself (http://www.nature.com/nature/journal/v437/n7058/i ndex.html ) (subscription required) indicates they've done more than that. By including the effects of coevolution - where one position in the protein mutates in concert with another to maintain optimal contacts - they generate a substantially better algorithm for manufacturing particular folds. (ie: 25% success in achieving folding versus 0% for conservation alone. 60% presence of wild-type function in the 'designed' proteins.)
Interesting, but I'm suprised it made it into Nature. (OK then, jealous...)Every time someone uses experiments like this to "disprove creationism" (as if this is something that can be proven or disproven in the first place), they ignore the common feature to all of these tests:
The scientist.
All this "proves" is that it is possible for an intelligent being to combine elements to create something more complicated.
You're saying this experiment shows all of this around us "proves" it could all happen on its own?
The rules for how DNA encodes protiens have been known since before I was born. The evolutionary mapping of how the genes coding different protiens duplicated and evolved into new structures is fairly easy to map out (give or take a brute force algorithm that runs in double-factorial-time to search through all evolutionary trees looking for the one that minimizes the number of mutations required along the way).
So this group has calculated the most likely common ancestor of the gene that now codes for a whole family of protiens, encoded the solution in real DNA, stuck it into bacteria and shown that it actually does produce a protien that they have been able to isolate the actual protien so that they can explore what it does/did.
(the term "articial protien" seems very odd to read - before I think it through, it sounds as though its hinting there is something mystical to "natural" protiens untouched by humans)
While I realize this news seems fascinating to some individuals, it is not something so entirely new that people in Computational Biology would consider it groundbreaking. Using the computer algorithms to generate new gene sequences is actually just a matter of running the gene finding algorithms you used to find these genes backwards (in fact, many people have been testing their gene finding algorithms by using their old algorithms to generate pseudo test sets). The only thing new about this paper is that people actually went forward and experimentally validated their results. An interesting find, however, the end result does not provide a huge leap to science.
Now, if people are really interesting in some neat ways of reengineering genes back onto themselves, then they should take a look at some of the work being done with synthetic circuits. The beauty of synthetic circuits is that since you already know how the genes will function, it's just a matter of setting the circuit up in the fashion that you want so that it produces the end result that you want. There really is no limit to what you can do with synthetic circuits (of course, researchers have a long way to go before they master and understand all the regulatory mechanisms). For example (and these are all very theoretical examples): building a cell circuit to release a drug into a body in a very time released fashion (and perhaps autonomously renewing, for example, building a circuit to release insulin into the body given the sugar level of the individual), designing a circuit to recognize and destroy tumors (or perhaps an even simpler form of designing a circuit to recognize and fluorescently label tumor cells in the body helping in removal/early detection). Of course, one could also build quite malicious synthetic circuits as well. For example, a circuit that would aggregate to the wall of the heart and, after a certain number of other cells accumulate, triggering a signal to all the malicious cells and destroy the heart in unison.
The other nice advantage of synthetic circuits is that the more we learn out regulatory mechanisms in species, the more we can use them for synthetic circuits. The more we use them for synthetic circuits, the more we understand about how exactly the underlying mechanism works (what causes them to break, how do they deal with differing toxic environments, etc). It creates a nice feedback loop with the progression of science.
There will come a day where it will be useful to generate new DNA/Proteins in combination with synthetic circuits, but, as noted in a previous post, we don't understand the relationship between protein sequence and structure/function enough for it to be a viable option (and this is just with how the protein folds, we haven't even gotten in to the problem of gene regulatory structures-- multiple gene splicing, chromosome structure elements, binding motifs, translational regulation, etc). In fact, this area is something we probably want to venture into as it provides us with an even finer control over the rate constants for synthetic circuits. But for now, the generation of randomly generated genes based on prior genes will go overlooked for quite some time.
I do not pretend to be an expert, or actually know anything about the subject and to promote Slashdot standards on information digesting, didn't actually read the article. But if one creates proteins that are exactly as nature would manufacture the real thing, why are they called artificial? They are the real thing!
PDFs of our papers, and Java code implementing 4 different correlated mutation algorithms including SCA, are at my web site:
http://www.afodor.net
The references are:
Anthony A. Fodor, Richard W. Aldrich. On Evolutionary Conservation of Thermodynamic Coupling in Proteins. JBC 279(18):19046-19050, 2004
John P. Dekker, Anthony Fodor, Richard Aldrich and Gary Yellen. A pertubation-based method for calculating explicit likelihood of evolutionary co-variance in multiple sequence alignments. Bioinformatics 20:1565-1572, 2004
Anthony A. Fodor and Richard W. Aldrich. Influence of Conservation on Calculations of Amino Acid Covariance in Multiple Sequence Alignments. Proteins 56(2): 211-221, 2004
The last paper contains a comparison between SCA and three other correlated mutation algorithms.
As I said, I haven't had a chance to look carefully or critically at the new papers. (It takes me a LONG time to read a paper critically :-> This Slashdot thread will be likely long archived before I finish thinking about these papers!). But this particular algorithm aside, people who are interested in bioinformatics and contact prediction may find the math behind the correlated mutation algorithms interesting.
Anthony
Email: anthony.fodor(remove this and put in an at symbol)gmail.com
http://www.afodor.net/
"Is life so dear, or peace so sweet, as to be purchased at the price of chains and slavery?" - Patrick Henry
The creationist/ID policy is to avoid facing unknowns by passing the buck onto a designer. In the current example, just because something appears elegant and simple to some person, it does'nt mean that it could not have naturally occured.
I don't think that's quite correct. My understanding is that ID examines a result using statistical or logical tools to see if it could have occurred by chance. It's not a subjective test. A statistical abberration indicates some outside influence. A collection of pre-existing conditions that all must be met at once (and not a step at a time) indicates some outside influence.
Glib oversimplified statements about ID will only come back to haunt you someday when we realize those ID guys were onto something - even if they don't quite have it all figured out yet.
Why are there symbiant relationships? It allows for division of labor, essentially. The genetic load of one organism after symbiosis does not have to take care of these certain task that the other is taking care of. Most of the cells contained in your body are not actually yours. The majority of cells in the body are bacteria living in your intestine which each produce proteins which help with digestion. If our DNA had to encode for every one of those digestive and metabolic proteins that are actually used in digestion, we would be selected against compared to an organism that could make more efficient use of its DNA.
Diversity also leads to a sort of long term stability. If there are different ways to obtain resources, the ecosystem as a whole can adapt to environmental changes far more gracefully.
I'll never make that mistake again, reading the experts' opinions. - Feynman
Yeah, exactly. Natural selection is pass/fail, there are no As, Bs, etc. :-)
Furry cows moo and decompress.
Not true species, is the only testable (ie scientific, ie non-arbitrary) classification. All the rest are arbitrary. A species is defined as a population of organisms which have reproductive isolating mechanisms which prevent the production of a viable (including sexual) offspring. ie two organisms are members of a different species when they are unable to reproduce and/or produce a viable offspring.
I dunno, GP might be right. The fact that a Scientist *can* drop a rock and have it hit the ground, probably means that rocks *cannot* fall off mountains without the aid of some Intelligent Dropper.
Censorship is telling a man he can't have a steak just because a baby can't chew it. --Mark Twain
"This tells us nothing that isn't already known. Of COURSE proteins with related functions share specific patterns of amino acid selection rules or they wouldn't work""
This cannot be assumed; while logical on the surface, that's like saying that porpoises and sharks must be from the same family because they both swim in the ocean and have the same body shape. The scientific method DEMANDS that a hypothesis like that is tested.
"He sounds like he's talking to a little kid. "
In terms of protein chemistry, he probably is talking to to little kids. This article is basically a publicity press release, and is not intended for the scientific community -- their papers in Nature are intended for the scientific community.
and how did it do this?
While you're at it, why don't you ask that any articles about scientific research include the entirety of the published paper(s)? What is important to the casual reader (which is the intended audience) is not the how, but the what and the why.
"One can only assume that these guys chose proteins that don't undergo post-translational modification."
No, one can't assume that. One must read their published papers before leaping to conclusions. Think a bit: if the DNA was expressed in the bacteria, could the proteins not undergo the same post-translational modification as naturally occurring proteins? The research conducted did not test to see what happened during expression, it just tested the form and functionality of the end result, as far as I can tell FTA.
You shouldn't get all worked up because an article intended for the general public isn't detailed enough for you. Many people wouldn't even bother to read TFA (even if they could understand it) if it was written in anything other than plain english.
Someone like you, who wants better information (which is a good thing) should go down to the library at the end of the month and read the published papers in Nature.
"Trolls they were, but filled with the evil will of their master: a fell race..." -- J.R.R. Tolkien on Olog-hai
I'm not sure how simple the rules are has anything to do with the chance that a particular thing arises. A simple rule could be "A standard earth protein must have 37 XYZ chains" - short, sweet, and simple. But how does now knowing that there must be 37 XYZ chains to get a standard protein now mean that before it was known it was simple? Especially when there are more potential combinations than there are atoms in the universe.
I stick by my earlier assertion that simple rules are only simple after they are discovered, not before when the future is unknown, and there are an incredible number of possibilities.
I've been on enbrel for six months now to treat my psoriasis (think lizard man from lepar island, not just crusty elbows). It's a protein that I have to inject twice a week since taking it orally would end with my body digesting the meds. I could see genetically engineering a bacteria that could live in the intestine and produce the medicine. That would be awesome.
Disconnect your television. Do your own research. Draw your own conclusions. They're probably lying. Don't be a sheep.
Since protein engineering is my field of study, for the benefit of the /. crowd (and my karma) I'll fill in the gaping holes left in the New Scientist article, and give you a little more background on the Nature paper. Because the writeup on /. is a perfect example of "scientific telephone": a semi-interesting result gets written up into a paper, which once it's been through several layers of editors suddenly seems like a major breakthrough.
The Nature paper isn't a breakthrough. It's not even really a major advance. Scientists in my field have been creating artificial proteins for five to ten years now. And yes, even some of them designed completely from scratch (though they're really simple; nothing as complex as, say, ATP synthase) instead of just taking a known fold pattern, known as a "motif." The "WW domain" (domain, in protein parlance, is a small, independent structure within a much larger protein---think of it like a module within the kernel or Apache) is a common fold in hundreds of different proteins. Basically, they analyzed the sequences of all of these WW domains, and figured out which positions were meaningful. It's kinda like reading through some code in a programming language you don't know, and figuring out which lines are comments and which lines are actual compilable code. This group found that the number of interesting positions is small, that they could identify them just from the amino acide sequence instead of having to mess with the whole complicated 3D structure of the domain, and that if they put together a protein with the meaningful amino acids intact and the non-meaningful positions randomized, then in many cases they could still get a pretty decent protein (in terms of structural similarity to the "natural" protein) out of it. Most of the paper is devoted to showing via various methods that they did get a pretty decent protein.
So what does this mean for me, assuming that this paper is absolutely correct (which I admit is a little hard for me to determine with one quick reading, given that I'm just a first-year grad student)? It means that the number of meaningful amino acids in a protein (at least in terms of overall structure) is pretty low, and that they can be identified without knowing what the full 3D structure is. This is good, because for a lot of proteins, the 3D structure is difficult to get. However, they picked an easy target: a small domain where there are over 100 unique sequences known. We'll see how well this method holds up with longer domains and fewer unique sequences. The S/N ratio won't be nearly as good.
Scientists don't use evolution because they're out to win some sort of ideological battle. They use it because it works. They want to understand how an amino acid becomes a protein. Evolution helps explain the process. Creationism and ID don't explain the process at all. They just say two things: (a) god made them; and (b) stop asking questions.
If evolution didn't help further science, it would be abandoned in favor of whatever did. But it works, which is why scientists rely on it and why teachers need to explain it in science classrooms.