Would one of the science experts explain what the Rosetta does exactly in a way that your average non scientificly minded computer geek, such as most of us here, can understand.

I have read into this site and there is a great deal I don't understand at all.

Can this be presented to us in lessons that are easy to digest.

Thank you ever so much

Todd Mack

---

Join TEAM MAUISUN.ORG

Start on the home page and look at the Research overview in the About section.

---

imagine a joint like your knee, you can try to twist it in thousands of ways, however, the less pain you feel the more likely it is that it will actually bend that way.
in proteins, you have lots of sections that can bend around different ways. the best way/the most likely way it folds is the one that uses least energy (=kinda pain).
ok, i admit thats not the best analogy but maybe the idea came across nevertheless.
rosetta tries to improve methods to predict how it will fold without observing it first.

---

--
Florian
www.domplatz1.de

So we are not actually finding a cure so much as we are trying to prep these protiens for the scientists to mess with correct?

---

If you cant say something nice, say something vague.

And if you can't think of something vague, then shut the hell up! =)

there are other projects that use rosetta/robetta for specific protein structure search (e.g. for cures, but they are not necessarily upen dc projects), this projects basicly tries to improve the search method. if i remember correct, baker lab as such is doing specific searches as well (i think they got funds for some HIV related research just recently) and we might get some of that to crunch later, too, but its not the top priority of this boinc project.

---

--
Florian
www.domplatz1.de

So we are not actually finding a cure so much as we are trying to prep these protiens for the scientists to mess with correct?

My understanding is that we're helping them produce a 'black box' where you can put in an amino acid sequence on one end and get the correct shape out the other. . . or you can put in the shape and get the sequence out.

The first direction allows you to take a sequence coded by DNA and see which protein it makes, the other direction lets you build a protein that attaches to something (say AIDS) in a useful way (let's your immune system attack it).

edit: stupid username, forgot to log back in :)

---

So we are not actually finding a cure so much as we are trying to prep these protiens for the scientists to mess with correct?

My understanding is that we're helping them produce a 'black box' where you can put in an amino acid sequence on one end and get the correct shape out the other. . . or you can put in the shape and get the sequence out.

The first direction allows you to take a sequence coded by DNA and see which protein it makes, the other direction lets you build a protein that attaches to something (say AIDS) in a useful way (let's your immune system attack it).

edit: stupid username, forgot to log back in :)

With the human genome project, scientists have now estimated that humans have about 20k genes.. This number is a problem because before the human genome project most scientists would have bet their houses that there is somewhere in the 100k range of genes.

Most scientists intuitively know that If you put a human clones (genetically same) in blenders under different stimulii (like anger, happiness.. etc) and did protein gels, you would have FAR more that 20k different proteins.
This concept challenges the central dogma in molecular biology..
DNA-->transription-->RNA-->translation-->AA-->protein

In other words there is more and more evidence that like in virus and bacteria one gene may code for more than one protein.

Proteins fold, bind, and link to do all sorts of things that science, "proteomics", is trying to understand.


This project is useing CPU power to find and predict the correct protein shape and folding pattern. Once perfected having a computer program that can give a scientist the correct shape of a protein in a matter or minutes will help speed proteomics up to a hyper-warp level.

This new power in itself will speed up the research for cures for diseases dramatically.



I haven't read any of the literature on this yet, so I could be way off.. If so I apologize!!

---

Secret team meetings and the sharing of 3.2Terabytes of free software -->HERE!... Don't spy, we don't like spies!

Well that is a whole lot different that FaD isn't it. Cool concept but funky all at the same time.

Thank you all.

---

If you cant say something nice, say something vague.

And if you can't think of something vague, then shut the hell up! =)

---

Would one of the science experts explain what the Rosetta does exactly in a way that your average non scientificly minded computer geek, such as most of us here, can understand.

I have read into this site and there is a great deal I don't understand at all.

Can this be presented to us in lessons that are easy to digest.

Thank you ever so much

Todd Mack

Yet another attempt at a simple explanation. Hopefully this will work hand-in-hand with some of the other explanations, to help clear things up.

First things first. Proteins are nothing other than long chains of amino acids. There's twenty different ones used in all (?) life on earth.

If you've heard of DNA and/or the genome project, that refers to the blueprint that's used by your body to build all the proteins needed.

OK, so your body has just finished making a protein, let's say it's for one of the cells in a muscle in your arm. For that protein to correctly do it's job, it has to fold, and it folds into a very specific shape, determined by the sequence of amino acids. All well and good, except for one minor problem. A ball park estimate for the TOTAL number of possible shapes the protein could assume is about 10^300 - yes, 10 to the 300th power.

My reasoning for this rather large number? Hopefully David Baker will set me straight if I'm wrong ..... There could be about 500 rotatable bonds in a typical protein, we have that on good faith from the Rosetta team. Assuming each rotatable bond can take on one of four individual orientations (this is a guess on my part), that's about 4^500, which works out to about 10^300.

What Rosetta does is to take the sequence of amino acids that make up the protein, and with no other knowledge try to determine what shape it'll fold into.

Why do we care about this?

If we can crack the problem of determining the shape a protein will fold into, purely by means of a computer simulation, it opens the doors to research on diseases that relate to incorrect folding (Alzhiemers?), and it also paves the way to making it much easier to craft "custom" proteins. These, in turn, have a lot of other potentially useful applications.

---

First things first. Proteins are nothing other than long chains of amino acids. There's twenty different ones used in all (?) life on earth.

20 are the ones used in the human body (and as far as i know in most other life forms), there are some more exotic (edit: exotic from our point of view) ones making there appearance in e.g. bacterias

---

--
Florian
www.domplatz1.de

Wow, you guys are really good at explaining this stuff to each other. The knee (joint) analogy is really good, and the overview by Aegion below is pretty much covers it. Let me add a few words.

A major premise of molecular and structural biology is that Structure == Function, ie the function of a protein is determined by it's 3-dimensional structure. So if we know the 3D structure of a protein, we can (usually) deduce its function. Also, if we know the 3D structure, we can design inhibitors to gum it up, if we want to inhibit a troublemaking protein. For example, an oncogene that has gone amok and is causing cancer; or a bacterial or viral protein that we would like to shut down. Or to take an example from the comment below, we could potentialy design a protein to disrupt the plaques formed by misfolded Alzheimer's proteins. Or, theoretically, we could design a protein to turn on a gene that is silenced in people with certain genetic diseases. The possibilities are endless.

Now the race is on to see whether the 3D structures of all proteins of biological and medical interest will be determined computationally, or experimentally. As you might know, a number of groups world-wide are developing higher through-put methods of experimental structure prediction, and they have solved many more structures than have been predicted computationally. But my money is on computational prediction, because once the program is working properly, there is no way experimental structure solving can be as fast or efficient as computational structure prediction.

A few other notes/trivia:

20 are the ones used in the human body

Actually, even in the human body there are more than 20 amino acids that are used; but only 20 are encoded in the genetic blueprint of our DNA. The rest are made by modifying the others post-translationally, or through other pathways than translation.

...you would have FAR more that 20k different proteins.
This concept challenges the central dogma in molecular biology..
DNA-->transription-->RNA-->translation-->AA-->protein

It is correct that there are many more proteins in the "proteome" (the full complement of proteins expressed by a cell, or organism) than there are genes. However, while it was surprising just how few genes there are in humans (actually about 30K), it is not surprising that there are more proteins than genes. A process called alternative splicing (which acts at the RNA level) generates different isoforms of a gene - ie more than one protein per gene. Random weird biological fact: A fruit fly gene known as Dscam has >30 000 protein isoforms, ie this one gene encodes more proteins than the number of genes in the rest of the fruit fly genome! Anyway, point is, more proteins doesn't contradict the central dogma.

So we are not actually finding a cure so much as we are trying to prep these protiens for the scientists to mess with correct?

Yes, the road from basic research to disease cure is a long one. It has to be, because we really need to know what we are doing before we try anything on humans! But here's a potential path that could be followed: 1) Computationally model a protein designed to inhibit a disease causing pathway; 2) Test out the inhibition in vitro ie in a test tube; 3) test out the inhibition in vivo ie in an animal model of the disease; 4) Go to clinical trials; 5) Cure the disease!
The bad news is this pathway from bench to clinic is a long one; the good news is each of the steps is getting faster and more effective as thousands of scientists are working on each step. And now all of you are involved at step 1, so thanks for that!

---

...you would have FAR more that 20k different proteins.
This concept challenges the central dogma in molecular biology..
DNA-->transription-->RNA-->translation-->AA-->protein

It is correct that there are many more proteins in the "proteome" (the full complement of proteins expressed by a cell, or organism) than there are genes. However, while it was surprising just how few genes there are in humans (actually about 30K), it is not surprising that there are more proteins than genes. A process called alternative splicing (which acts at the RNA level) generates different isoforms of a gene - ie more than one protein per gene. Random weird biological fact: A fruit fly gene known as Dscam has >30 000 protein isoforms, ie this one gene encodes more proteins than the number of genes in the rest of the fruit fly genome! Anyway, point is, more proteins doesn't contradict the central dogma.

[/quote]


Good points, but I was trying to make it simple for non-scientists.
I think a lot of us here could have gone into alternative splicing and Drosophila.. but that really might go beyond what TM was wanting.

---

Secret team meetings and the sharing of 3.2Terabytes of free software -->HERE!... Don't spy, we don't like spies!

http://users.rcn.com/jkimball.ma.ultranet/BiologyPages/T/Transcription.html

here is a link that discuss about others splicing, but it is not exactly the most non-technical approach, i see if i find something more detailed/easy to read later today
one part where splicing is essential for human surviving is the immune response, we could never produce all the various antibodies if it was not for splicing and recombination of the different "partial" coded pieces.

---

--
Florian
www.domplatz1.de

Would one of the science experts explain what the Rosetta does exactly in a way that your average non scientificly minded computer geek, such as most of us here, can understand.

I have read into this site and there is a great deal I don't understand at all.

Can this be presented to us in lessons that are easy to digest.

well you all lost me in a load of scientific gobbledygook there........

when i needed to explain to others what i was doing at find-a-drug, it went a bit like this :-

well the scientists have created a crystal structure of a target........and we compare a libruary of 12 billion known small drug molecule shapes to a site on the target looking for a "best fit" ........ the computers mathematically compare shapes and give a score to the best fits.......best score should in theory be a cure.

i'm a dustman by trade (trash man).........so a simple (no more than 100 words) similar to the above would be usefull for those of us who want to do something usefull with the cpu's.....but not get lost in dna strands and amino acids and other things i don't understand.

---

46 years dc so far

join team FaDbeens
join us

so a simple (no more than 100 words) similar to the above would be usefull for those of us who want to do something usefull with the cpu's.....but not get lost in dna strands and amino acids and other things i don't understand.

Great idea! Maybe start out with an easy-to-understand few sentences that anyone could use at a party to explain briefly what we are up to. And then provide more and more complex detail like the previous posts so that one could "dig into" as much detail of the project as each person wishes! :)

---

Regards,
Bob P.

The bad news is this pathway from bench to clinic is a long one; the good news is each of the steps is getting faster and more effective as thousands of scientists are working on each step. And now all of you are involved at step 1, so thanks for that!

Thank you very much for your graciousness! That is a part of what typifies this project, and helps make it a lot of fun!

---

Regards,
Bob P.

Thank you everyone for the many different takes on this project. I was starting to feel a little goosey about doing this as compared to FaD meaning that I could explain FaD a whole lot easier to people.

You have all set my mind at ease and I want to thank you all for taking the time to do so.

Now how about we bust these jobs out and find a cure damnit!

---

If you cant say something nice, say something vague.

And if you can't think of something vague, then shut the hell up! =)

i'm a dustman by trade (trash man).........so a simple (no more than 100 words) similar to the above would be usefull for those of us who want to do something usefull with the cpu's.....but not get lost in dna strands and amino acids and other things i don't understand.

WHAT we're doing, in 30 words or less .....

Every protein in your body has to fold to a particular shape to do it's job. We're trying to figure what that shape is, using a computer simulation

WHY we're doing it, in 20 words or less .....

If we can crack this problem, it paves the way to helping cure a lot of diseases.

---

I said:

Every protein in your body has to fold to a particular shape to do it's job. We're trying to figure what that shape is, using a computer simulation

In hindsight, "simulation" is a bad word to use there, we're not simulating anything. I'd change that last phrase to read: ..., using a massive grid of computers.

---