How do I take screenshots without a paper background?
How do I import my own metadata?
How do I import my own PDB files?
How do I align my SVD axis to something meaningful?

How do I take screenshots without a paper background?

In the very first screen in the RoPE app, click on the Menu option in the top right hand corner. Click Options and set the background to white.

How do I import my own metadata?

Some metadata is generated within RoPE, e.g. single point mutations between the sequence of a structure and the reference sequence. It is also possible to download some additional metadata from the Protein Data Bank if the names of the models are PDB codes.

If you have custom metadata which you would like to import, this needs to be added as a CSV file with a specific format. Here are some examples:

model,dose (MGy)



Important notes:

  • The CSV file must have the file extension .csv.
  • Only use ASCII characters.
  • There must be at least one column called 'model' or 'molecule' for the CSV to be recognised as a metadata file.
  • Metadata can be numerical (as above). This allows you to draw line series or vary colour according to the numerical value. It can also be text (e.g. "oxidised", "reduced") for changing the display icon.
  • Line series will connect all values in a column in numerical order. This means that if you want two separate lines for two separate series, they must be entered in separate columns, like in the final example above, which prepares time(A) and time(B) for the A and B chains, respectively.
  • If you have a model column, this will match to the model name with which you have defined the structure. This information will be displayed for all molecules in this model. If you use a molecule column, this is in the format {model_name}_{chain_ids}, and would match the names given to the display icons when you hover over them. e.g. 3tm7_CD or 1pt1_B. This will only apply to the molecule specified, and not other molecules belonging to the same model.

    If you decide to export from spreadsheet software such as Excel or Numbers, start your first column in cell A1 (i.e. top left corner) and then export as CSV. If you have any problems, open the file in a text editor to try and see how it differs from the example above.

    To import this on the Web app, click 'Add' on the File Menu page, navigate to the CSV file, and upload it. To import this onto the Mac app, place the CSV file in your project directory and restart the application. Navigate to the File Menu.

    If all goes well, you will see the 'metadata information' icon appear to the right of the CSV file listed in the File Menu. If not, check your CSV formatting and make sure you have a 'model' or 'molecule' column on the first line.

    If you click on this file, you will get a summary of the metadata headers, and an option to add to the database. Click on 'Add to database' and your custom metadata will become available to RoPE.

    How do I import my own PDBs?

    Although the RoPE tutorials focuses on using data downloaded from PDB-redo, it is possible to do this analysis with your own PDBs. Please note that if you want to compare your own PDBs with those downloaded from PDB-redo, it would be prudent to pass your structures through the PDB-redo server first in order to use the same refinement protocol. Otherwise there is often an offset between differently processed groups of PDBs.

    The first step is to load these PDB files into RoPE. On the web app, this can be done with the Menu > Load function (or Add in the File Manager). On the Mac app, you must put the PDB files in the project directory before launching the program. These will be assimilated into the File Manager on launch.

    The new files will then behave like any other PDB files in the program. Auto-modelling these files will use the filename as the model name, and unless this matches a PDB code, you will not be able to fetch automatic metadata from the Protein Data Bank. However, you can import your own metadata to do so instead.

    There may be instances (such as custom proteins) where the Uniprot ID does not exist. It is still possible to make an entity from the sequence generated by the atoms in the PDB file. First, go to the models from the home page and click on Add model.... Next to Initialising file click Choose... and choose the PDB with the most complete modelled sequence. Enter Chain assignment using the right arrow.

    This shows that the entity of the chain is not assigned. By clicking on (not assigned) you are able to create an entity from the calculated sequence. You can also check the sequence by clicking on 341 res (or similar) on the right hand side.

    Once you have created the entity from the sequence, you can use automodelling to fill in the rest of the models from your own PDB.

    How do I align my SVD axis to something meaningful?

    The mathematical procedure of singular value decomposition does not necessarily align with your biological meaning. If you have two sources of biological variation with similar variance, this will result in having similar matrix eigenvalues. In this case, determining the eigenvectors becomes more volatile, and may end up on some combination of your two biological sources of variance.

    In order to see the default axes of the singular value decomposition, right click one of the molecules and click 'set as reference'. This will generate three arrows corresponding to each axis.

    In order to realign to some source of metadata, first, create a rule to colour according to some scheme. For example, click on rules, followed by (+) new rule. Choose a rule of type vary colour and Choose... a metadata class corresponding to numerical values (e.g. resolution). Return to the RoPE space, right click one of the axes, and click match colour.

    The (negative of the) correlation coefficient will be reported in a message; click to dismiss. This realignment is carried out in only the three axes of the SVD on display. You can now right click and explore this axis as a set of torsion angle variations applied to the protein.