Hello,
I have a question about the robetta fragment-file. what does the numbers mean?!
1ofd A 1068 S L -163.692 159.068 168.636 -0.177 4.773 84.262 3 0.000 P 1 F 1
1ofd A 1069 P L -59.041 138.502 179.354 -0.177 4.773 84.262 3 0.000 P 1 F 1
1ofd A 1070 L L -65.398 -35.892 176.967 -0.177 4.773 84.262 3 0.000 P 1 F 1
Are the first free values angles?!
I have the same understanding problems with the silent.out file.
1 L 0.000 115.760 -11.712 1.458 0.000 0.000 174.910 -16.319 0.000 0.000 S_00000001
2 L -66.362 -101.587 -182.586 2.717 0.969 2.285 30.184 -36.782 28.708 0.000 S_00000001
3 E -174.094 75.957 170.068 -0.100 1.839 4.685 177.349 164.795 179.226 -178.826 S_00000001
4 E -59.667 160.921 168.430 -0.296 5.284 6.266 -64.416 0.000 0.000 0.000 S_00000001
What mean the values?!
Regarding the fragment file, see https://www.rosettacommons.org/node/3510 and https://www.rosettacommons.org/node/2254 -- https://www.rosettacommons.org/comment/2079 is particularly illuminating, as it was written by the person who wrote the most recent version of the fragment picker for Rosetta.
On the silent file, those lines in particular are for a "protein" formatted silent file. See https://www.rosettacommons.org/docs/latest/silent-file.html
{residue pose number} {secondary structure code} {phi} {psi} {omega} {CA x} {CA y} {CA z} {chi1} {chi2} {chi3} {chi4} {tag}
Note that in the protein formatted silent file you only get the main dihedrals - distances and (three-center) angles are assumed to be "ideal".
Thank you very much.
I have now another question. What do Rosetta exectly with this fragments?! Through the amino acid sequence rosetta must link the fragments in the order of this sequence. Who Rosetta selected this fragments?! After the hole sequence is setup, rosetta chance the given angles to find the konformation with the lowest energy or is my imagination wrong?!
Fragments are used by Rosetta to sample backbone conformations. For example, in the standard ab initio folding run, many of the early stages are a Monte Carlo procedure where fragments are inserted into the polypeptide chain. (By "fragment insertion" we mean that Rosetta reads off the backbone dihedral angles from a fragment, and then changes the backbone dihedrals of the modeled protein to match. We "insert" the conformation of the fragment into the protein.) The fragment file is already laid out by position - each fragment in the fragment file already corresponds to a given section of the modeled protein - that matching is what occurs during the fragment picking stage. Once fragment picking is done, and you have an actual fragments file, there doesn't need to be any sequence matching, or assembly - each fragment has it's location, and each location has a set of fragments which corresponds to it. See [Simons KT, Kooperberg C, Huang E, Baker, D. (1997) Assembly of protein tertiary structures from fragments with similar local sequences using simulate anealing and Bayesian scoring functions. J Mol Biol 268:209-25.] for details
Other fragment use cases, like loop remodeling, are similar. For loop remodeling, you just insert fragments in the region of the loop, rather than the whole protein. The correspondence between positions in the modeled protein and fragments in the fragment file is the same, though.
As far as energies, most protocols are using a Monte Carlo procedure. Roughly, a random position is picked, and a random fragment for that position is selected. That randomly picked fragment is inserted and optimized slightly, and the energy evaluated. Based on the energy, the insertion is either accepted or rejected (See http://en.wikipedia.org/wiki/Metropolis%E2%80%93Hastings_algorithm). In either case, a new random position/fragment is selected and the cycle repeated many times. Repeated many times, most structures tend to fall into lower energy states. The fragment insertion is just a way of sampling biologically plausible backbone conformations for the protein.
Thank you very much.