ddG backbone movement

Member Site Forums Rosetta 3 Rosetta 3 – General ddG backbone movement

Viewing 1 reply thread
  • Author
    Posts
    • #1423
      Anonymous

        Hello,
        I am using Rosetta 3.3, ddg_monomer application. I am trying to reproduce protocol 19 from Table I (Kellog, 2010) but it looks like there is no backbone minimization at all. In Fig. 3 there is a visible backbone movement but in my case it behaves like in supplementary material Fig. 1, where no backbone movement is observed.

        my setting is: ddg_monomer.static.linuxgccrelease -in:file:s minimize_with_cst.XYZ.pdb -ddg::weight_file soft_rep -ddg::iterations 20 -ddg::dump_pdbs true -ddg::mut_file output_path.mut -ddg::local_opt_only false -ddg::min_cst true -ddg::mean true -ddg::min false -ignore_unrecognized_res -fa_max_dis 9.0 -ddg::dump_pdbs true -ddg::min true -ddg::ramp_repulsive true -ddg::minimization_scorefunction standard -ddg::minimization_patch score12

        Could you help me to incorporate backbone minimization (or flexibility) into my protokol?
        Dave

      • #7871
        Anonymous

          First off, you repeat a number of options. For example, you have -ddg:min twice, once false and once true. The latter should override the former, but for best results I’d only specify it once.

          Your core problem is that you haven’t specified ‘-ddg:sc_min_only’. By default it’s true, and if you want more than just sidechain minimization (i.e. you want the backbone to move) you need to explicitly set it false.

          (By the way, I’d recommend taking a look at the Kellogg et al. supplemental information – it gives example commandlines for most of the protocols in the table.)

        • #7873
          Anonymous

            Thanks for quick answer. Adding the -ddg:sc_min_only was one of the problems. I was using also wrong weight file (soft_rep instead of soft_rep_design). So supplemental information was realy usefull.
            Thanks a lot, best regards
            Dave

        Viewing 1 reply thread
        • You must be logged in to reply to this topic.