Back To Multistage Rosetta Scripts
This script was used to relax 30,000 abinitio results. All 30,000 structues went through one repeat of FastRelax. The 7,500 (25%) best results survived and went through the second repeat. After that, the 1,875 (25%) best results went through the final 2 repeats. This was roughly 3 times faster than running 30,000 through all 4 repeats using traditional rosetta scripts
<JobDefinitionFile>
<Job>
<Input>
<PDB listfile="my_pdblist"/>
</Input>
</Job>
<Common>
<SCOREFXNS>
<ScoreFunction name="common_sfxn" weights="ref2015.wts"/>
</SCOREFXNS>
<TASKOPERATIONS>
<InitializeFromCommandline name="ifc"/>
<IncludeCurrent name="ic"/>
<ExtraRotamersGeneric ex1="true" ex2="true" name="ex1ex2"/>
</TASKOPERATIONS>
<FILTERS>
<ScoreType name="sfxn_filter" score_type="total_score" scorefxn="common_sfxn" threshold="999999"/>
</FILTERS>
<MOVERS>
<FastRelax disable_design="true" name="relax_1" repeats="1" scorefxn="common_sfxn" task_operations="ifc,ic,ex1ex2"/>
<FastRelax disable_design="true" name="relax_2" repeats="2" scorefxn="common_sfxn" task_operations="ifc,ic,ex1ex2"/>
<SwitchResidueTypeSetMover name="to_fa" set="fa_standard"/>
</MOVERS>
<PROTOCOLS>
<Stage num_runs_per_input_struct="1" total_num_results_to_keep="7500">
<Add mover_name="to_fa"/>
<Add mover_name="relax_1"/>
<Sort filter_name="sfxn_filter"/>
</Stage>
<Stage num_runs_per_input_struct="1" total_num_results_to_keep="1875">
<Add mover_name="relax_1"/>
<Sort filter_name="sfxn_filter"/>
</Stage>
<Stage num_runs_per_input_struct="1" total_num_results_to_keep="1875">
<Add mover_name="relax_2"/>
<Sort filter_name="sfxn_filter"/>
</Stage>
</PROTOCOLS>
</Common>
</JobDefinitionFile>