Output repeats alignment in QuatSymm resutls

[AntoniyaAleksandrova]

Is it possible to add an option to output the final alignment between repeats detected by QuatSymm and used for the calculation of the reported TM-score and RMSD (just as is done in CE-Symm, for example)?

[lafita]

Hi @AntoniyaAleksandrova, if you are referring to the quaternary symmetry detector, an alignment is only done for pseudosymmetry, otherwise sequences are assumed to be identical and so the alignment is the identity.

The RMSD and TM-score reported are scores referring to the deviations from perfect symmetry, so they are an average of the rigid superposition of all possible permutations of the entire protein complex in 3D. So for example, the RMSD for a D2 symmetry complex would be the average of the RMSD from a superposition of the complex against a two-fold rotation through one axis, and a two-fold rotation through the second axis.

[AntoniyaAleksandrova]

Hi Aleix,

Thanks for adding the FASTA output option for the chains. I guess my question is the following. Given what QuatSymm currently outputs, how can one use this information to plot the symmetry axes and figure out the transformations that are related to it? Obviously, it is possible, since in the PDB database, the QuatSymm results are displayed with the corresponding symmetry axes and repeats (and this information is passed on to JMOL). However, I need to obtain this information and it's unclear to me how to do so.

For example, for the simple case of 3UX4 where there is a C6 symmetry detected, QuatSymm produces a FASTA alignment that has the chains ordered in the same way that they are in the reported Subunits list: [A, F, E, D, C, B]. However, to obtain the C6 symmetry, one needs to match A -> B -> F -> D -> E -> C cyclically in this order, which is not consistent with the order given in the FASTA or the report. So it is unclear to me how to produce the C6 axis from the QuatSymm results without figuring out the matches and transformations from scratch myself.
By contrast, CE-Symm clearly reports this information in the axes file, where it outlines both which repeat follows which around the axis and what the axis is in terms of two points that are on it.

Thanks again for your help!

[lafita]

Hi @AntoniyaAleksandrova , thanks for your question. I agree this would be a nice feature to add, and it should be relatively straightforward, but can't commit to making the changes soon. All the information on symmetry groups and axis is available from the QuatSymmetryResults object.

If you want to explore the code yourself have a look at the BioJava tutorial and Demo:

• https://github.com/biojava/biojava-tutorial/blob/master/structure/symmetry.md
• https://github.com/biojava/biojava/blob/master/biojava-structure/src/main/java/demo/DemoSymmetry.java