The 3DIANA-WS at a glance3DIANA is a web based environment designed to integrate bioinformatics-like information for the analysis of protein interactions and quaternary structure modellling. The first version of the application introduces the use of molecular interaction data to evaluate potential binding domains between proteins. Protein domains are defined according the Pfam classification  and calculated with the HMMER package . Domain-Domain Interaction scores are computed with DIMERO approach . Also Domain-Domain Interaction templates collected from 3DID  database are integrated in order to identify potential domain binding sites in protein structures and to provide a set of structural templates for protein docking.
Uploading Structure FilesThe application provides one webform to upload structure files (in PDB format) and and a second form to define the subunits. In this way the suer can define the structure of one proteins using multiple files, this is very useful when the structure of a protein has been solved in diferent PDB entries.
Figure 1 Upload files form.
Figure 1 shows the web form used to upload PDB structure files to the server. Clicking the 'ADD File' button creates a new file entry; however, no empty files or file entreies can be submited. The 'Submit' button will upload the selected files and the next form will be opened in the browser.
Defining Protein SubunitsOnce the desired structure files are uploaded the user will need to define the subbunits combining the different structres and chains conatined in the files. Figure 2 shows the form that is used to define the subunits.
Figure 2 Defining Subunits.
Clicking the 'Add Subunit' button (A) will create a new field to define a new subunit. The subunits will be used to group the different structures and chains submitted in the previous step and define the proteins for the subsequent analysis. In the subunit field the user can add new regions defined by the submitted structures. To add a new region, the user can chose among the struture files uploaded (C) and the PDB chains contained in the files (D). Once the structre/chain has been selected the user can add the new region to the corresponding subunit by clicking the 'ADD' button (B). The selected region will appear in the subunit table (E). When all subunits are defined the information can be submitted to web server clicking the 'Submit' button. 3DIANA will generate a 'normalized' input structurue where the structure associated to Subunit 1 will be identified with Chain A, Subunit 2 Chain B, etc. The chain residues will be reindexed and their IDs will start at 1 then it is worth to notice that the order in which the different regions are incorporated to a particular subunit is important for the reindexing process.
Domain-Domain Binding AnalysisThe 'Domain-Domain Binding Analysis' tool is used to analyze the phisical interaction between the protein domains among the subunits. The protein-protein interaction (PPI) matrix window ('PPI matrix' Figure 3) displays an overall scoring description of the interactions between the protein subunits. Each element of the matrix shows the best scored interacting domain pair between the proteins. In green high confident prediction (HCP), in orange medium confident prediction (MCP), in red low confident prediction (LCP) and in gray no significant score (NS) as defined in DIMERO .
Figure 3 PPI matrix.
The DDI table windowWhen one of the PPI matrix elements is clicked the application will display the Domain-Domain Interaction (DDI) table in a new window (Figure 4). This table contains a detailed description of the phisical domain interaction between the selected proteins. For each interacting domain pair the DDI-table shows the location of the domains in their respective sequence, the predicted score (HCP in green, MCP in orange, LCP in red and NS in gray as defined in DIMERO) and if structural evidence of the DDI is available the 'Struct. Model' column is marked with a green tick. When a row of the 'DDI table' is clicked the main 3D viewer will highlight the interacting domains.
Figure 4 DDI table.
Clicking with the right mouse button on the rows of the DDI table the application will display the DDI menu. This menu contains the options to access the different sources of information related with the selected domain pair.
The 'Fetch PPI' option is used to retrieve the PPI data that is needed to compute the sIP and sNP metrics (see The PPI resume window) as defiend in DIMERO . The 'Pfam *' options link with the domain entries of the Pfam database. Finally, when structural information of the interaction between the selected domains is available, the '3D Model 3DID' options links with the corresponding entry of 3DID database .
Domain-Domain Interaction AnalysisThe 'Domain-Domain Interaction Analysis' tool is used to analyze the potential interaction between the domains of two selected subunits based on DIMERO scores. In this scenario 3DIANA scores all possible combination of domain pairs independently of the 3D conformation. Figure 5 dsiplays how this information is organized, the user can select among the different subunits using the selection buttons (Figgure 5A). The, the DIMERO score information is diplayed in a table for all possible domain pairs, also the 'Struct. Mode' column indicates if structural evidence of the DDI is available with a green tick or not with a red corss. The radio button located in the left site of the table (Figure 5C) can be clicked to map the domains in the protein structures and sequences (Figure 5B). Finally, the 'Fetch PPI' (Figure 5D) button is used to retrive the interactomic data uased to compute DIMERO scores, a more detailed documentation on this panel can be found in DIMERO help pages.
Figure 5 DDI sequence analysis.
Domain Binding Sites AnalysisThe 'Domain Binding Site Analysis' tool is used to explore the known the binding regions and interacting partners of a particular domain with other domains. This information is compiled from 3DID database. The select button (Figure 6A) allows to choose among the differnt subunits, then the domains of the selected protein are displayed in a table where the radio buttons located on the left site allow to select any of them (Figure 6B). Finally, the binding site search can be filtered by sequence identity (Figure 6C) and only the information of those domains that their sequence identity is greater than the selected threshold will be displayed.
Figure 6 Domain Binding Sites Search.
When the 'Fetch Binding Sites' buttonn is clicked 3DIANA will show a new window describing the binnding site regions, see next.
Domain Binding Sites InterfacesThe 'Domain Binding Sites' window display the known binding sites of the selected domain at sequence level. The binding sites information are clustered in two levels: first level cluster gorups those binding sites that are used by diferent domain partners with different conformmations (Figure 7B) and second level groups the structures that shares a similar 3D conformation of their inetrfaces (Figure 8B). The structure of the interactions contained in a particular cluster can be deisplayed clickng the Jsmol icon, then a new window and 3D viewer will be dsiplayed to explore all possible conformation of the selected cluster (see next section). The cluster panel (Figure 7B) represents the interacting amino acids over the domain sequence with red segments, this innformation is useful to a quick view of the binding site location. When these charts are clicked the interacting amminoacids will be highlighted on the domain sequnce panel (Figure 7D) and the main 3D viewer. Also, the clsuter panel displays the maximum sequence identity between the domain under study and the domains containedd in a particular clsuter (Figure 7C).
Figure 7 Domain Binding Sites Interfaces.
When a particular structure of a cluster is selected (Figure 8C) the domain sequnce panel is updated and the sequence alignment between the domain under study and the domain of the particular selected structure is displayed in the domain sequnce panel (Figure 7D). The interacting amino acids are highlighted in red color in the domain sequnce and also on the residues of the main 3D viewer (Figure 8A).
Domain Binding Sites 3D ViewerThe structure binding sites contained in the different clsuter can be explored clicking the Jsmol icon (Figure 7A) in the 'Domain Binding Sites Interafces' window. Then, a new window is opened (Figure 8B) displaying a new 3D viewer and a table with all the structures compiled in the selected cluster. This table identifies the PDB structures that were used to collect the specific bindign information contained in the cluster; however, only the structures with sequence identity higher than the selected score are displayed (see Figure 6C).
Figure 8 Domain Binding Sites Viewer.
Finally, the user can fetch the structure of a particular interaction to the main viewer and compare the structure of the domain under study and the domain of the particular interaction, also the binding partner is collected. Clicking in the 'Fetch Structure' button the structure of the selected intercation will be algnied with the domain under study and displyed in the main viewer (Figure 9).
Figure 9 Domain Binding Site Alignment.
Domain-Domain Template Docking3DIANA provides a tool to preform template-based docking when structural information for two domains of the submited proteins are available. The 'Domain-Domain Template Docking' tool allows to perform protein dokcing based on solved DDI structures. Figure 10 shows the 'Domain-Domain Template Search' window, the user can select among the different submited proteins using the select buttons (Figure 10A) and all their domain pairs for which structural information is available are displayed in a table.
Figure 10 Domain-Domain Template Search.
The user can filter the available DDI templtes to explore only those that have a sequence identity greater than a selected threshold. In this case the sequence identity is determined by the minimum when both selected domains are comapred with a particular tamplate. Finally, clicking the 'Fetch DDI Templates' botton will open the 'Domain-Domain Template Interface' window to explore the different available templates, see next.
Domain-Domain Template InterfacesThe 'Domain-Domain Template Interfaces' window (Figure 11) describes the DDI templates and their interfaces, it consists of two panels the DDI interafces panel (Figure 11A) that provides a quick overview of the interacting residues and the sequences panel (Figure 11D) that display the sequence of the protein domains selected for docking and a possible template when selected (next section, Figure 12C). The DDI templates are grouped using a single level clustering where each cluster contains the templates with a similar 3D interface conformation (Figure 12B). Clickng the Jsmol icon (Figure 11C) will open the a new window and 3D viewer to explore the structure of the templates contained in the selected cluster (Figure 12B).
Figure 11 Domain-Domain Template Interfaces.
When a particular template is selected (Figure 12C) the domain sequnce panels are updated and the sequence alignment between the selected domain pair to perform docking and the domains of the particular selected template are displayed (Figure 11D). The interacting amino acids are highlighted in red and green colors in the domain sequences and also on the residues of the main 3D viewer (Figure 12A).
Domain-Domain Template 3D ViewerThe structure of the templates can be explored clicking the Jsmol icon (Figure 11A) in the 'Domain-Domain Template Interfaces' window. Then, a new window is opened (Figure 12B) displaying a new 3D viewer and a table with all the templates compiled in the selected cluster. This table identifies the PDB structures and the domains of the templates contained in the selected clsuter; however, only the structures with sequence identity higher than the selected score are displayed (see Figure 10B).
Figure 12 Domain-Domain Template Viewer.
Finally, the user can perform template-based docking using a particular templateclicking in the 'Dock Structures' button. Then, the selected domains will be structurally aligned to the selected template, also the template will be fetched and displayed in the main viewer (Figure 13). The initial structures are not modified but a copy of the green domain subunit (Figure 10A) is created and placed/aligend with the red domain protein (Figure 13).
Figure 13 Domain-Domain Template Docking Model.
3DIANA Chimera Plug-in is a component to interafce 3DIANA WS with Chimera and replace the web jsmol viewer. To install the plug-in, download the .tar.gz file and extract its content in the <CHIMERA_PATH>/share folder. To run now 3DIANA in Chimera go to Tools menu, Utilities, click 3DIANA and the web server will open the 3DIANA Chimera starting page.
References Finn R.D., et al. (2014) Pfam: the protein families database, Nucleic acids research, 42, D222-230.
 Johnson L.S., et al. (2010) Hidden Markov model speed heuristic and iterative HMM search procedure, BMC bioinformatics, 11, 431.
 Segura J., et al. (2015) Using neighborhood cohesiveness to infer interactions between protein domains, Bioinformatics, 10.1093/bioinformatics/btv188.
 Mosca R., et al. (2014) 3did: a catalog of domain-based interactions of known three-dimensional structure, Nucleic acids research, 42, D374-379.