Member Site › Forums › Rosetta 3 › Rosetta 3 – Applications › score to use for picking top docking solutions?
- This topic has 3 replies, 2 voices, and was last updated 13 years, 2 months ago by
Anonymous.
-
AuthorPosts
-
-
June 14, 2011 at 1:47 pm #945
Anonymous
Hi,
I have done protein-protein docking using rosetta 3.2 to generate 60,000 structures. I’m a bit confused about the next steps.
1) Which score, the total score or I_sc, to use for picking solutions for further consideration? Which of these are referred to in publications?
2) one of the older webpages I found describes three methods for picking solutions (http://graylab.jhu.edu/~mdaily/tutorial/blind_main.html). Are these still followed as a general procedure by the community?
Any help would be appreciated.
thanks.
-
June 14, 2011 at 4:43 pm #5762
Anonymous
We typically look at I_sc. I_sc is the “interface score,” which is the total score of the complex minus the sum of the scores of the individual docking partners. Values of about -5 to – 10 typically indicate a good decoy. It is important to also look at the total score, as a clash or other problem in one of the partners can distort the I_sc significantly. These problems are most common when incorporating flexible backbones into docking.
That web page is still what we recommend using, however the scripts used haven’t been ported to work with Rosetta 3 yet, so they may need to be tweaked a bit. If you run into specific problems, please post them. Finally, totally blind predictions haven’t been thoroughly tested in Rosetta 3, so please keep that in mind as you procede.
-
June 24, 2011 at 5:34 pm #5814
Anonymous
The best way to do this seems to be to download a copy of Rosetta++ and get the scripts from the rosetta_scripts/docking directory. I don’t have experience using the cluster app, but it seems like it’s worth a try.
-
June 15, 2011 at 8:35 am #5767
Anonymous
Hi Weitzner,
Many thanks for this and the specific note on blind docking.
I’m not sure where to find those scripts referred in that web page (and I’m not good with scripts either!). I’ve done sorting with the I-sc values. What I’d like to be able to do is extract the top 200 pdb files from my output directory into a new directory and cluster them. Is there a way to do this? or can the cluster app do this in one go . i.e. pick the top 200 based on I-sc and output the corresponding pdbs to a new dir?
Thanks for the help.
-
-
AuthorPosts
- You must be logged in to reply to this topic.