SwissModel

SwissModel is a server for automated comparative modelling tasks. It is an extremely powerful and flexible resource which we are fortunate to be able to access free of charge.

Go to the SwissModel site: http://swissmodel.expasy.org/ and have a look at the page and the options which it provides.

Click on the Modelling menu item at the top of the screen where you will see four main options:

In this practical we are just going to use the automated First Approach mode to create a model and get a feel for the process and what it consists of. Detailed documentation for the server is available at http://swissmodel.expasy.org/docs/help

Using SwissModel in First Approach Mode

The sequence we are interested in, is a myoglobin sequence from bloodfluke.

>sp|O77003|GLB_BIOGL Globin (Myoglobin)                      
SLSDADKKALDASWKKLTAGADGKKNAGINLVLWMFANVPNMRAQFSKFNANQSDDALKG 
DAEFIKQVNVIVAALDGLLQSVNNPGQLQANLDKLAKSHVNLKIGLEFFGPLQQNIHSFI 
ESALGVGAGSDEPKAWGNLIAAFNETLKKA                               

We could use the server in first approach mode for this protein either by copying and pasting the sequence into the text box or simply by entering the accession code (O77003).

However, the server would take some time to respond and a lot of requests at once would only exacerbate this. We will therefore look at the results without actually needing to submit the job. Since the user's involvement in the process is minimal this should not compromise the usefulness of this process.

The format of the output has changed very recently so if you run this with SwissModel itself rather than looking at the precomputed results, things will look a little different.

In addition, the structure of a new homologue has recently been solved and deposited in the PDB. Consequently, the potential templates listed will also have changed!

The server returns an email with information about the submission:

Read the Acknowledgement email

Once the run has finished, the results and various useful information about the run are made available on the web.

In particular, you can learn which template structures were chosen on which to base your model, how similar they are to the target sequence and which regions of your sequence could be modelled. In addition, any warnings generated during the alignment and modelling process will be displayed. If the process fails to produce a model then this file provides information as to why.

Visit a copy of the page returned by the server and examine clearly the information it contains:

Read the Server results

Look carefully at the log information. First look at the section entitled Template Selection Log.

Record how many putative templates were initially found.

Record how many of these were not rejected because the sequence identity with the target sequence was too low and were therefore used to create the 'Graphical Template Match Overview'.

Templates are reported as extended PDB ("ePDB") codes - simply the PDB accession code for the structure with the letter indicating which protein chain is used added to the end, for example 1aujA is chain A from the bovine trypsin structure with the PDB accession code 1auj.

A single template was selected by the algorithm for final use in modelling the protein. Record the 'ePDB' code of the template structure used.

Now look at the section entitled Modelling Log. Record the reported final energy of the model.

 
Continue