I am a beginner in rosetta software and need help to build params file

Member Site Forums Rosetta 3 Rosetta 3 – General I am a beginner in rosetta software and need help to build params file

Viewing 2 reply threads
  • Author
    • #831

        I am a beginner in Rosetta. I would like to use match to define a binding site with constraints. But my ligand that I want to define is a metal. So I define my metal in a pdb file as:
        HETATM 3835 FE HEM 1 17.140 3.115 15.066 1.00 14.14 FE3+

        Then I create the .mol file by using avogrado software and I compile in rosetta by using molfile_to_params.py
        But it isnot working. So I compil the demo that rosetta propose and it is ok. Apparently the problem is my ligand because in the .mol file I have only one line.

        So my question is: Can I define in match a metal as a ligand?
        If not What is the part in rosetta which permit me to insert a binding site with constraints in a scaffold (which defines) and to have a metal as ligand?

        thanks a lot

      • #5198

          The Rosetta database has params files for some metal ions already (see minirosetta_database/chemical/residue_type_sets/fa_standard/residue_types/metal_ions), including both iron III (residue name FE) and iron II (residue name FE2). The “trick” to make monoatomic ions work is to put in virtual atoms such that the metal ion can be oriented in space. molfile_to_params.py isn’t really set up to do this, so if you want to make a metal ion parameter file, your best bet is to copy one of the existing files and tweak it appropriately. Note that if you want to use a residue that’s already in the database, you don’t have to make a new params file, or even specify the existing one with -extra_res_fa. If you use the appropriate residue name, Rosetta should be able to find the params file in the database. (So for matching, a “-match:lig_name FE” without the -extra_res_fa should be sufficient.)

          I note that your residue is listed as heme – note that if you’re trying to place a metal ion in combination with another small molecule (an iron coordinated heme, for example), your best bet is to treat the entire complex as a single ligand. All you need to do is add “dummy” single bonds so that the entire “residue” is connected by bonds, and then pass that mol file through molfile_to_params.py. (Note that you don’t need the virtual atoms then, as the other atoms in the residue serve to orient it.)

          Finally, keep in mind that the parametrization of metal atoms in Rosetta is not as developed as the standard protein atoms (most importantly, orbital geometry isn’t considered). Matching is purely geometric, so you should be fine, but I recommend you be a bit skeptical of the results coming out of energy minimization or design runs. That is, check that the results make sense chemically.


        • #5215

            Thanks for answering me. Now I have my param file good, but my constraint file (.cst) is not good,
            When I compile, just after reading my .cst file, Rosetta said me “Segmentation fault”.
            For example, if I want to define a binding between my metal and a histidine of my scaffold, is it correct to define like this?:
            TEMPLATE:: ATOM_MAP: 1 atom_name: FE
            TEMPLATE:: ATOM_MAP: 1 residue4: Fe3p

            TEMPLATE:: ATOM_MAP: 2 atom_name: HIS,
            TEMPLATE:: ATOM_MAP: 2 residue4: N

            CONSTRAINT:: distanceAB: 2.03 0.07 100.00 1 0
            CONSTRAINT:: angle_A: 102.67 4.00 100.00 360.00 1
            CONSTRAINT:: angle_B: 153.54 4.00 50.00 360.00 1
            CONSTRAINT:: torsion_A: -146.7 4.00 50.00 360.00 2
            CONSTRAINT:: torsion_B: -8.03 4.00 25.00 180.00 4
            CONSTRAINT:: torsion_AB: -6.43 6.00 0.00 180.00 5

            Thanks again


          • #5217

              It doesn’t look like the enzdes/matcher constraint file documentation is all that well linked on the released manual, but http://www.rosettacommons.org/manuals/archive/rosetta3.2.1_user_guide/app_match_enzdes_cstfile.html is a direct link.

              First off, it looks like you have the residue names and the atom names switched. Secondly, I don’t believe “residue4” is an allowed term. It should be either “residue1” (for one or more standard amino acids as one letter codes) or “residue3” (for natural and unnatural amino acid three letter code – “FE” counts as a three letter code – you may need a space after it to satisfy the parser, though). Keep in mind that “atom_name” takes three atom names the first is the one to which the distance is measured, the second helps define the angle, and the third helps define the torsion. Also, note that the atom name “N” refers to the backbone nitrogen of the amino acid – if you want the sidechain nitrogens, you’ll need to use “ND1” or “NE2” for the atom names of the delta or epsilon nitrogen, respectively. Or if you don’t care which one it is, you can specify “atom_type Nhis” for a deprotonated nitrogen or “atom_type Ntrp” for a protonated one (the matcher should consider multiple histidine protonation tautomers). “atom_type” will automatically figure out what the other two atoms should be based on the fold tree (the connectivity in the residue params file).

              Finally, as you have a monoatomic atom for your “A” residue, you don’t really need to have extensive sampling of the angle_A, torsion_A or even the torsion_AB parameters. With the current settings, you’re going to spend a lot of time just spinning the virtual atoms all over the place. I’d recommend setting the periodicity of the A-side angles/torsions to 360 and the sampling parameter to 0 – this will minimize the time wasted placing the virtual atoms in different positions. Setting the distanceAB and B-side angles and torsions however you like should be sufficient to get the geometry to the histadine how you want it.

          Viewing 2 reply threads
          • You must be logged in to reply to this topic.