Yet another off the wall question.

When looking at the "results" page I understand RMSD, at least I think I do.

But I have no idea how "score" is arrived at. Looking at the scores posted I assume that the further into negative territory, the better,

Also, at the top of the page is the title "Active Work Units" - I assume that this means that other users are still processing on this work unit.

Which raises the question: when the dust settles and all processing on a work unit is complete, where can we go so see how we did?

Thanks

RMSD is a measurement of how similar a protein you have calculated is to its original known structure. The lower the number the closer the match to the original, with 0 being a perfect match. Because RMSD relies on knowing the original structure it is a completely useless measurement when you are doing a blind test and trying to calculate an unknown model from scratch. However the thinking behind using RMSD on known models is if you are able to design a process that finds the perfect model with 0 RMSD for a known structure then the same process should be able to find the perfect model for an unknown structure.


"Score" is a measurement of the natural energy of the protein structure. As one of my Science teachers explained to me many years ago, molecules are very "lazy" as they like to form themselves with the least amount of energy possible. In other words, they form the easiest possible configuration available. One of the key ideas with Rosetta is that the model with the lowest energy you can find is going to be the real structure. Much of the Rosetta@home research is aimed at trying to find a process that calculates the lowest energy structure accurately and with the least amount of time and computer power.

---

Which raises the question: when the dust settles and all processing on a work unit is complete, where can we go so see how we did?

Thanks

Each workunit is actually only a small part of a bigger unit that was split into smaller ones so our pc's can handle it in an efficient manner. At some projects there are only 4 workunits to one bigger unit, on others there could be hundreds. And then each workunit can be sent out to several people all at once, again depending on the project, and then when all the parts come back, and are validated, they are recombined into their original unit, which itself could be part of a much larger unit. So basically there is no way to know until the work gets published, or they tell us in some other way.

Murasaki explained things in terms even I could understand ...

Thanks for the information!

Each workunit is actually only a small part of a bigger unit that was split into smaller ones so our pc's can handle it in an efficient manner.

Is it possible to include something in each work unit graphics to show us what part from the original structure we are working on? I hope this makes senses

---

Have a crunching good day!!

Each workunit is actually only a small part of a bigger unit that was split into smaller ones so our pc's can handle it in an efficient manner.

Is it possible to include something in each work unit graphics to show us what part from the original structure we are working on? I hope this makes senses

Normally you will be working on exactly the same structure as everyone else. I believe what mikey is trying to say is that your single work unit generates a number of "decoys" or individual results and that all the decoys from all participants in that batch are gathered together into a combined set of results. Other batches of work units may be sent out with slight variations on the original batch to see if changing the method of calculation has any effect.

Once all the decoys from all the work units from all the related batches are gathered together the scientists can then analyse the data to see which method worked best and if there was any unusual behaviour that they need to take account of in the next set of tests.

Unfortunately we can't tell the impact of the decoys we submit except when Bakerlab publish a scientific paper or one of the team make an announcement in the Rosetta@home Science forum. The most regular source of information on the impact of the project is Dr Baker's Journal.

---