Hi All,

wow - it's been a long time since i posted here. So - what's been going on ?
Those of you following the forums will have seen us labouring away to make minirosetta more stable. The latest result of these effords is minirosetta_1.54
which we have been running hard on BOINC over the last 3 weeks and that has proven to be quite a bit more reliable then the prvious version. There will be a few more minor updates, but all in all its time to focus on science again !

Yeah! So we currently have a lot of stuff going on. We currently have about 2.5 Million WUs (!) queued up here - that's gonna generate quite a large amount of data over the next few weeks!

For homology modelling we are finally crunching to find out some vital answers about all the different methods we use to establish which ones work best in which situations as well as to optimize their parameters.
We're about to launch some more investigations into those algorithms that align unknown with know sequences to establish how to best make use of the huge amount of protein structural data that scientist have already collected - there'll be another 1.5M or so workunits going out soon!

We are also optimizing the energy function and are finding some curious things about the area of conformational space directly surrounding the native structure. We're currently crunching a massive amount of sampling to get very finegrained. In some cases Rosetta's lowest energy structure deviates a little from the true structure and we're investigating if this is due to xtal packing. If so then maybe rosetta is actually giving a more representative answer of what the protein looks like in solution then suggested by the crystal. Very curious.

Further there is more docking going on, as well as a number of investigations into how to improve sampling for abinitio structure prediction by varying the energycomponents.

All in all this a super-exciting time for us & we'll be posting interesting developments here !

Thanks for crunching !

Mike

---

http://beautifulproteins.blogspot.com/
http://www.miketyka.com/

Thank you for the post!

I think you guys should have a special section on the front page with scientific updates like this one. It is too easy to miss them on the forum.

And, probably a link to http://depts.washington.edu/bakerpg/ would be nice idea. Or whole section on the front page with name of publication and a links to PDFs. Look at other projects like MilkyWay@home, QMC@home, Spinhenge@home(just a link) etc.

For homology modelling we are finally crunching to find out some vital answers about all the different methods we use to establish which ones work best in which situations as well as to optimize their parameters.
We're about to launch some more investigations into those algorithms that align unknown with know sequences to establish how to best make use of the huge amount of protein structural data that scientist have already collected - there'll be another 1.5M or so workunits going out soon!

Looks like you stopped to name homology methods benchmark workunits using hombench_ prefix you have mentioned before.

---

TestPilot, AKA Administrator

I think you guys should have a special section on the front page with scientific updates like this one. It is too easy to miss them on the forum.

I guess, that's a very good idea to involve public in the research.