WebSTAR3D: a web server for RNA 3D structure alignment

Try an example.

First Structure:
PDB code     or upload a file:     chain id:

input format:

Second Structure:
PDB code     or upload a file:     chain id:

input format:

Render With:  

Gap Open: Gap Extension:
Match Score: Mismatch Score:
RMSD Cutoff: Minimum Stack Size:
Update 7/11/2017: WebSTAR3D no longer employs MC-Annotate for base pairing annotation.

To calculate the 3D structure alignment of two RNA molecules, input for each structure:
  • A valid four letter Protein Data Bank (PDB) identification code associated with a molecule of interest (e.g. 3cw5 or 3rg5) OR a valid PDB or mmCIF formatted file.
  • A valid chain identifier.
The default algorithm parameters of gap open/extension cost, match/mismatch score, RMSD cutoff, and minimum stack size have been empirically chosen to provide the best result. However, they can be modified in the form accessible by clicking "Expand Parameters".

To render the alignment: If your browser supports Java applets (see How do I enable Java in my web browser?), Jmol is recommended for the best performance. While JSmol takes significantly longer to render, it is universally compatible.

Before aligning structures, WebSTAR3D preprocesses PDBs with base pairing annotation using DSSR and pseudo-knot removal using RemovePseudoknots. To improve runtime, all parsable PDBs on the 1.89 release of non-redundant lists of RNA stuctures under a 4.0A resolution cutoff have been preprocessed. When entering PDB codes into the form, the form offers suggestions for the codes corresponding to these preprocessed PDBs, along with suggestions for their available chain identifiers. The cache of preprocessed PDBs will periodically update with new releases.
WebSTAR3D globally aligns RNA structures with a two-step strategy: first, it finds the consensus of stacks between the structures using 2D topology and 3D geometry; then, it uses this information to guide the alignment of loop regions. The stack and loop alignments are then combined into the final result. This powerful approach avoids complex and expensive secondary structure comparison and base matching, making the running time feasible for the comparison of large RNAs. Yet, it matches the prediction accuracy of state-of-the-art tools.

The STAR3D algorithm and WebSTAR3D are described in the following papers:
  1. Ge, Ping, and Shaojie Zhang. "STAR3D: a stack-based RNA 3D structural alignment tool." Nucleic acids research 43.20 (2015): e137-e137.
  2. Holzhauser, Erwin, Ping Ge, and Shaojie Zhang. "WebSTAR3D: a web server for RNA 3D structural alignment." Bioinformatics (2016): btw502.
The source code for a Java implementation of the STAR3D algorithm is available at: http://genome.ucf.edu/STAR3D/.
For bug reports, comments, or suggestions please contact Shaojie Zhang, Erwin Holzhauser, or Ping Ge.