Back to Filters page.


Compare a set of homologous but structurally heterogeneous PDBs to a template PDB and find structurally highly conserved sites that can serve as stems for splicing segments.

<StemFinder name="(&string)" from_res="(1&int)" to_res="(pose.total_residue()&int)" rmsd="(0.7&float)" stems_on_sse="(false&bool)" stems_are_neighbors="(true&bool)" neighbor_distance="(4.0&float)" neighbor_separation="(10&int)" filenames="(&comma-separated list of pdb file names)"/>
  • from_res, to_res: template positions (in rosetta numbering) in which to search for stems. (positions out of range will be ignored).
  • rmsd: cutoff for the average rmsd between a given position in the template and all of the closest positions in the homologs.
  • stems_on_sse: demand that in each of the homologs the candidate stems are on 2ary structural elements. This isn't a good idea, b/c DSSP is a bit noisy
  • stems_are_neighbors: should we eliminate stems that are farther than neighbor_distance from one another?
  • neighbor_distance: minimal atomic distance between any pair of atoms on each of the residues.
  • neighbor_separation: minimal aa separation between candidate stem sites
  • filenames: PDB structures that are well aligned to the template. Use align or cealign.

See also