GPCRM web service

A web service to predict structures of G-protein-coupled receptors (GPCRs) using advanced homology modeling tools: profile-profile alignment, multiple structural templates, and Z coordinate-based filtering of resulting models. Two distinct loop modeling techniques are employed - Modeller and Rosetta - and the proper scoring is used to score the final ranking list of GPCR models.

Quick Path.
This option is of the first choice (if there is no high similarity between a target and a template). The loops are modeled with Rosetta. The obtained models are of good quality which can be used for ligand docking and for MD simulations.

Long Path.
This is the longest and the most precise option. The loops are modeled with Rosetta and the precision can be increase further by using Rosetta slow modeling. The obtained models are of better quality than from Quick Path but for most cases the Quick Path is enough.

High similarity.
This fast option can be used only when there is a high similarity between a target and a template sequence. Only in this option all loops between transmembrane helices are preserved in shape from the template since they are not modeled by Rosetta (this option in unavailable here). There is also no possibility for advanced (manual) mode - only Auto mode. N- and C-termini are obtained in extended conformations if they are not present in the template structure and can be cut off.

Filling the input data entry.

The selection of templates is based on sequence similarity between query and templates calculated using multiple sequence alignment of a query sequence and all GPCRs sequences whose 3D structures are available in PDB. In the current version of the service we used a set of templates structures presented in the Templates section . If you mark the lysozyme option, then during modeling a set of templates is narrowed to those which has lysozyme version.

Here type your name or nickname (maximum number of characters: 100). It is optional.

We suggest to type your e-mail address here, but it is optional. The link to your results will be mailed to you. Alternatively, you can copy and save that link from the submission page (in case no e-mail is provided).

Here type the name of your temporary task for convenient identification of your results. For example, let's type 'CB2 cannabinoid receptor - inactive' since this is the receptor we are going to model.

Allowed format: fasta or plain sequence. To obtain the sequence of CB2 cannabinoid receptor go to http://www.uniprot.org/ and search for 'CB2'. Then click on P34972 at the very left of 'CNR2_Human' entry, scroll down until you find a sequence and click 'FASTA' button. Either copy-paste the CB2 receptor sequence in the box provided, or save FASTA sequence file to disc and then click "Choose File" button below to upload the sequence file.

To improve accuracy of the alignment and model building provide the suggested disulfide bridge between the extracellular loop 2 and the trans-membrane helix TM3. In our example (CB2) the disulfide bridge is not located between the EC2 and TM3 so those fields will be left empty.

Choosing the right options.

Task Mode: Auto - all steps of the modeling procedure are processed without any User's intervention.
Advanced - a User can change: subset of templates used in the modeling (first step) and the query-templates alignment and loops which are remodeled (second step). Since this tutorial is for beginners we choose Task Mode: Auto

Set of Templates: We provide 2 subsets of templates: PDB structures with agonists (named: active ) and PDB structures with inverse agonists or antagonists (named: inactive). For the purpose of this tutorial we want an inactive CB2 receptor model, so in this field we choose 'Inactive'.

Lysozyme: This option was left for historical reasons only and will not be discussed here. For this tutorial leave the default setting of this parameter 'Do not add from template'.

Loops present in the model can be remodeled by Rosetta loop-modeling protocol.

'Run Query' button tarts the initially selected one of 3 main routes (Quick path, Long path, or High similarity), here in Auto mode.

This service uses cookies.