Note: This page is not a complete listing of all RosettaScripts conventions
RosettaScripts is becoming a significant tool in the Rosetta community, and most of the movers and filters in Rosetta already exist in the form of a RosettaScript parsable module. The lack of a standard RosettaScript XML file created a situation where each module (mover or filter) performed identical functions in different ways. This document was designed to correct this issue - a standard form will enable other application (pyRosetta, RosettaDiagrams) to work with RosettaScripts much more elegantly, and make it much easier to even hard-code.
Internally, Rosetta gives every residue or ligand in a structure a unique index, starting with 1. Most movers and filters allow users to specify residue indices using the Rosetta numbering. In addition, many allow users to specify residue indices using PDB numbering (an index and a chain letter, e.g.
47C), though this should be used with care since it assumes first that the input structure came from a PDB file, and second that numbering has not changed over the course of a protocol.
A new, experimental feature has recently been added to permit users to specify residue indices based on some reference state of a pose. For example, let's say we had a case in which we were going to import a structure, insert an unknown number of residues in a loop around position 35, and then mutate a residue at what used to be position 50, but which now could have any of a range of indices. Prior to inserting residues, a "snapshot" or reference pose can be stored in the pose using the StorePoseSnapshot mover, and after inserting residues, the user can say, "I want to mutate the residue that used to be residue 50". The syntax for this is
refpose(<refpose_name>,<index_in_refpose>). Additionally, relative positions can be specified. For example, a user could say, "I want to mutate the residue that's two residues before the residue that was residue 50". The syntax for this is
Currently, only the MutateResidue mover and the HbondsToResidue filter have been modified to take advantage of this new feature. However, movers can be modified very easily to respect reference poses. See this note in the development documentation for details on how this can be done.
All score function attributes should begin with scorefxn. In case 2 score functions are used (for example, low and high resolution) an underscore should be added: scorefxn_low, scorefxn_high.
A task operations attribute should be named task_operations, and will contain a comma separated list (without spaces) of all task operations.
There are several ways you can include other modules in your own module:
Use the attribute mover_name or filter_name within your protocol tag.
use the <Add mover_name=.../> or <Add filter_name=.../> tags. for example:
<Your_Mover some_attr="..."> <Add mover_name="looper"/> <Add filter_name="minimizer"/> </Your_Mover>
Please refrain from using logical operations within your tag, as it will only overpopulate the schema. If you want to enable logical operations that contain filters, please use the CompoundStatement Filter. For example, the following Code:
<SomeMover ... > <AND filter_name="filter1"/> <AND filter_name="filter2"/> </SomeMover>
<CompoundStatement name="compoundFilter" .. > <AND filter_name="filter2"/> <AND filter_name="filter1"/> </CompoundStatement> <SomeMover ... filter_name="compoundFilter"/>
<CompoundStatement name="compoundFilter" .. > <AND filter_name="filter2"/> <AND filter_name="filter1"/> </CompoundStatement> <SomeMover ... > <Add filter_name="compoundFilter"/> </SomeMover>