In the second step, the continuous fragments excised from the PDB templates are reassembled into full-length models by replica-exchange Monte Carlo simulations with the threading unaligned regions (mainly loops) built by ab initio modeling. In cases where no appropriate template is identified by LOMETS, I-TASSER will build the whole structures by ab initio modeling. The low free-energy states are identified by SPICKER through clustering the simulation decoys.

In the third step, the fragment assembly simulation is performed again starting from the SPICKER cluster centroids, where the spatial restrains collected from both the LOMETS templates and the PDB structures by TM-align are used to guide the simulations. The purpose of the second iteration is to remove the steric clash as well as to refine the global topology of the cluster centroids. The decoys generated in the second simulations are then clustered and the lowest energy structures are selected. The final full-atomic models are obtained by REMO which builds the atomic details from the selected I-TASSER decoys through the optimization of the hydrogen-bonding network (see Figure 1).

Figure 1. I-TASSER protocol for protein structure and function prediction.

For predicting the biological function of the protein (the last column at Figure 1), the I-TASSER server matches the predicted 3D models to the proteins in 3 independent libraries which consist of proteins of known enzyme classification (EC) number, gene ontology (GO) vocabulary, and ligand-binding sites. The final results of function predictions are deduced from the consensus of top structural matches with the function scores calculated based on the confidence score of the I-TASSER structural models, the structural similarity between model and templates as evaluated by TM-score, and the sequence identity in the structurally aligned regions [A similar approach to structure-based function annotation was proposed by Brylinski and Skolnick (PNAS 2008. 205:129) who tried to match the target structures on the threading templates. Here the I-TASSER server matches the target models on all template proteins in the libraries].

The I-TASSER server (as "Zhang-Server") participated in the Server Section of 7th (2006), 8th (2008), 9th (2010), and 10th CASPs (2012), and was ranked as the No 1 server in CASP7 and CASP8. In CASP9 and CASP10, I-TASSER server and QUARK (another server from our lab) were ranked as No 1 and No 2 servers, respectively. The detailed rank results can be seen here for CASP7, CASP8, CASP9, and CASP10. Figure 2 shows histograms of the Z-score of GDT-TS scores of all servers in CASP7 (68 servers), CASP8 (72 servers), CASP9 (81 servers), and CASP9 (72 servers).

Figure 2. Histogram of Z-scores of all server groups at CASP7, CASP8, CASP9 and CASP10.

Figure 3 is a summary of COFACTOR, a component of I-TASSER server, in the function prediction section of CASP9, where COFACTOR was registered as "I-TASSER_FUNCTION" and "Zhang" in the server and human prediction sections, respectively. The picture was taken from the presentation by the CASP9 assessor Dr. T Schwede, see http://predictioncenter.org/casp9/doc/presentations/CASP9_FN.pdf.

Figure 3. Mean MCC Z-scores of the best ten groups in the Function Prediction in CASP9.

• Up to five full-length atomic models (ranked based on cluster density)
• Estimated accuracy of the predicted models (including a confidence score of all models, and predicted TM-score and RMSD for the first model)
• GIF images of the predicted models
• Predicted secondary structures
• Predicted solvent accessibility
• Top 10 threading alignment from LOMETS
• Top 10 proteins in PDB which are structurally closest to the predicted models
• Predicted Enzyme Classification and the confidence score
• Predicted GO terms and the confidence score
• Predicted ligand-binding sites and the confidence score
• An image of the predicted ligand-binding sites

If users know some information about the structure of the modeled proteins, the information can be conveniently uploaded to the I-TASSER server. These information can significantly improve the quality of structural and function predictions.

The I-TASSER server currently accepts two types of user-specified restraints:

(1) inter-residue contant and distance restraints;
(2) template structures and template-target alignments.

• Assign contact/distance restraints: If you know what atom pairs should be in contact or in some distances, you can use this option to upload a text file including the contact and/or distance information of atom pairs.
  
• Specify template without alignment: If you want I-TASSER to use a specific PDB structure as a template, you can use this option specify the PDB structure. You only need to type in the PDBID:ChainID, e.g. 1wor:A without specifying the target-template alignments. If the chain information is not present in the PDB file, indicate the ChainID using "_". I-TASSER will first fetch the structure from the PDB library and then generate the target-template alignment based on our in-house alignment tool, MUSTER.
  
• Specify template without alignment: You can actually use any 3D structure as the template, which does not necessary exist in the PDB library. In this case, you can use this option to upload the 3D structure. This structure file must be in the standard PDB format. You do not need to input the target-template alignments. I-TASSER will generate target-template alignment based on our in-house alignment tool, MUSTER.
  
• Specify template with alignment: This option allows you (usually the advanced users) to specify both template structure and the target-template alignment.

I-TASSER needs templates to generate high-resolution structure predictions. In general, excluding close templates will decrease the quality of the I-TASSER modeling. However, users can exclude some templates from the I-TASSER template library for some special purposes (e.g. knowning some templates are different from target, or benchmark testing of the current algorithms).

The I-TASSER server accept two ways of template excludings:

• Exclude templates that are homologous to the query protein: The users can use this option to exclude templates from the I-TASSER template library, which are homologous to the query protein. The homology is defined based on the sequence identity cutoff, i.e. the number of identical residue between template and query divided by the total number of residues in the query sequence. For example, if you type "60%", I-TASSER will automatically exclude all templates which have a sequence identity >60% to the query protein. The minimum cutoff is set at 25% and all value below 25% will return as 25%.
• Exclude specific template proteins: This option allows users to upload a list of template structures that will be excluded from the I-TASSER template library. As the PDB library is redundant and same protein can exist as multiple entries, I-TASSER server will by default exclude the user-specified templates as well as all templates that have a sequence identity >90% to the specified templates. Users can also specify a different sequence identity cutoff, e.g. 70%, where I-TASSER will exclude all templates with a sequence identity >70% to specified template proteins.

  The format of the file should be "PDBID:ChainID %Sequence_Identity", e.g.

  1wor:A 70
  3mxu:A 80
  1zko:B 40
  OR
  1wor:A
  3mxu:A
  1zko:B

Currently, the major time consuming part in the I-TASSER protocol is the structural refinement assembly simulations. For those users who want a quicker reponse or those who do not need a refined models, we recommend them to use our LOMETS (meta-server) or MUSTER (single-server fold-recognition). Because these two servers do not attempt to refine the threading models, the response time is faster than the I-TASSER server.

• Y Zhang. I-TASSER server for protein 3D structure prediction. BMC Bioinformatics, vol 9, 40 (2008). (download the PDF file).
• A Roy, A Kucukural, Y Zhang. I-TASSER: a unified platform for automated protein structure and function prediction. Nature Protocols, 5: 725-738 (2010) (download the PDF file)