The input of COACH-D can either be the sequence or the structure of the query protein. Users can also submit a query ligand for docking.
As shown in Figure 1, the COACH-D workflow is as follows:
As shown in Figure 1, the COACH-D workflow is as follows:
(1) If the sequence of the query receptor (monomer only) is provided, the I-TASSER Suite or AlphaFold2 will be used to predict the receptor's structure.
(2.1) If the structure is a monomer, it will be processed by five individual methods (TM-SITE, S-SITE, COFACTOR, FINDSITE and ConCavity ) to predict the protein-ligand
binding pockets and residues. Four of these methods make predictions by searching the query structure/sequence through the Q-BioLiP and BioLiP template libraries.
(2.2) If the structure is a multimer, it will be processed by the quaternary-based method Q-SITE.
(3) Both template-searched ligands and user-submitted ligand are docked into the predicted binding pockets using the efficient docking algorithm AutoDock Vina.
(2.1) If the structure is a monomer, it will be processed by five individual methods (TM-SITE, S-SITE, COFACTOR, FINDSITE and ConCavity ) to predict the protein-ligand
binding pockets and residues. Four of these methods make predictions by searching the query structure/sequence through the Q-BioLiP and BioLiP template libraries.
(2.2) If the structure is a multimer, it will be processed by the quaternary-based method Q-SITE.
(3) Both template-searched ligands and user-submitted ligand are docked into the predicted binding pockets using the efficient docking algorithm AutoDock Vina.
Figure 1. The flowchart of the COACH-D algorithm.
Job submission guide:
(1) Input or upload a protein sequence in FASTA format(only supports single chain) or a protein structure in PDB, mmCIF format(supports single-chain & multi-chains).
(2) Provide the ligand file(optional).
(3) Choose whether to keep your results private.
(4) Submit.
Notice: Due to computing resource limitation, we now allow no more than 10 running/pending jobs per user at the same time.
(2) Provide the ligand file(optional).
(3) Choose whether to keep your results private.
(4) Submit.
Notice: Due to computing resource limitation, we now allow no more than 10 running/pending jobs per user at the same time.
Figure 2. The submission page of COACH-D.
There are separate download links for the following:
(1) The submitted protein
(2) The submitted ligand (optional)
(3) AZIP file containing the main results
(2) The submitted ligand (optional)
(3) A
Figure 3. The download link of COACH-D.
The COACH-D results are usually summarized in a web page that consists of three sections:
(1) The display frame of COACH-D results.
(2) The main results table of COACH-D.
(3) Summary of template ligands.
(1) The display frame of COACH-D results.
(2) The main results table of COACH-D.
(3) Summary of template ligands.
(1) The display frame of COACH-D results.
This section displays the predicted binding of five pockets to the ligand.
Predicted complex structure: Visualization of predicted complex structure.
Current complex Information: Detailed information of the predicted complex on the left side.
This section displays the predicted binding of five pockets to the ligand.
Predicted complex structure: Visualization of predicted complex structure.
Current complex Information: Detailed information of the predicted complex on the left side.
Figure 4. The display frame of COACH-D results.
(2) The main results table of COACH-D.
This section displays more details about prediction.
Rank : The predictions are ranked by the C-score.
Poseu: The ligand binding pose submitted by user.
Poset: The ligand binding pose obtained from the PDB template.
Cscore: The confidence score of the prediction. C-score ranges [0-1], where a higher score indicates a more reliable prediction.
Cluster Size : The number of protein template in a cluster.
P Template : The representative protein template in a cluster.
L Template: The representative ligand template in a cluster.
TM-score: The harmonic average of two TM-scores normalized by the query and the template, respectively.
Predicted binding residues: The consensus prediction based on spatial clustering of the predicted binding sites from individual methods.
Poseu: The ligand binding pose submitted by user.
Poset: The ligand binding pose obtained from the PDB template.
Cscore: The confidence score of the prediction. C-score ranges [0-1], where a higher score indicates a more reliable prediction.
Cluster Size : The number of protein template in a cluster.
P Template : The representative protein template in a cluster.
L Template: The representative ligand template in a cluster.
TM-score: The harmonic average of two TM-scores normalized by the query and the template, respectively.
Predicted binding residues: The consensus prediction based on spatial clustering of the predicted binding sites from individual methods.
Figure 5. The main results table of COACH-D.
(3) Summary of template ligands.
This section displays the ligand information from the table above, excluding the representative ligands listed in the "L Template" column.
Rank: The ligands are ranked by their frequencies in the cluster.
Ligand: Click on the ligand ID to view more information in Q-BioLiP database and click on the download button to get the corresponding SDF file.
Frequency: The total number of times the ligand occurs in the cluster.
PDB template: The PDB structures of the ligand in the cluster. At most 20 templates are listed for each ligand.
Visualization: Visualizaiton that can be rotated and zoomed in/out with the mouse (powered by 3Dmol).
Others: Other ligands in the cluster. The numbers in brackets are the frequencies. At most 100 ligands are listed.
Ligand: Click on the ligand ID to view more information in Q-BioLiP database and click on the download button to get the corresponding SDF file.
Frequency: The total number of times the ligand occurs in the cluster.
PDB template: The PDB structures of the ligand in the cluster. At most 20 templates are listed for each ligand.
Visualization: Visualizaiton that can be rotated and zoomed in/out with the mouse (powered by 3Dmol).
Others: Other ligands in the cluster. The numbers in brackets are the frequencies. At most 100 ligands are listed.
Figure 6. Summary table of template ligands.
Please cite the following article when you use the COACH-D server:
Q Wu, Z Peng, Y Zhang, J Yang, COACH-D: improved protein-ligand binding site prediction with refined ligand-binding poses through molecular docking, Nucleic Acids
Research, 46: W438–W442 (2018).
Q Wu, Z Peng, Y Zhang, J Yang, COACH-D: improved protein-ligand binding site prediction with refined ligand-binding poses through molecular docking, Nucleic Acids
Research, 46: W438–W442 (2018).
If you have more questions or comments about the server, please email yangjy@sdu.edu.cn.