Example dataset:

TSV File: E.coli-Bing Yang(2012) Reference: Yang B. et al. Nat. Methods (2012) (PDF)(HTML)


Video Tutorials




View Interactive Walk-through in Cytoscape.js:




Documentation and Frequently Asked Questions

Basic Information

Inputing Data and Starting Your Analysis

Protein Modeling and Docking

Network Analysis and Interactomes

Species Level Interactions


If you do not find your question here, please email XLinkDB at uw dot edu with your questions/comments.

Basic Information


  1. What operating system supports XLinkDB?

    Currently, XLinkDB runs on both Windows and OS X. Linux's web browser java plugins will not run XLinkDB.

  2. What web browsers should I use?

    XLinkDB runs best on the latest Firefox but can be run on the latest Safari and Chrome web browsers. Internet Explorer is not recommended because the script runs slowly.

  3. What organisms does XLinkDB support?

    Proteins from any organism are supported, but to ascertain information from protein interaction reference databases, currently only E.coli and H.sapiens are supported. More organisms will be incorporated into XLinkDB later.

  4. What protein interaction reference databases are used in XLinkDB?

    Primarily, XLinkDB uses data from IntAct (H.sapiens) or EciD (E.coli).

  5. What structural viewer can I use to visualize a protein model?

    XLinkDB currently has two ways to view protein structures and cross-linked residues: (1) JSmol and (2) NGL Viewer. If you want to edit and save structures that you have viewed on XLinkDB, use NGL Viewer which has many tools and resources for on-the-fly edits of your protein of interest.

  6. What should I do if I found a bug or I have a brilliant suggestion?

    Join us in our Google discussion group

Return to top of page.

Inputing Data and Starting Your Analysis


  1. What input file format should I use?

    A tab delimited text file. Arrange into six columns as follows:
    Peptide A | Protein A | Labeled position A || Peptide B | Protein B | Labeled position B

    The labeled position is the position of the labeled site within the peptide starting from 0. Please do not use a table header/column names.

    A template file is available at the top of this page for your reference [E.coli-Bing Yang(2012)].

    Note: Please rename your files before uploading to XLinkDB so that the file names include only letters and numbers.

  2. Why should I use Uniprot identifiers?

    XLinkDB 2.0 mines both Uniprot and the Protein Databank (PDB) to identify protein sequences and empirical protein structures that already exist. Uniprot accession numbers are used to query for these data.

  3. Does it matter what cross-linker was used to generate my data?

    No. The data input is independent of the cross-linker used to generate the data.

  4. I made a private table, how can I see it?

    On the Homepage under "Choose dataset", write in your table name. If you are unsure of your table name, email XLinkDB@uw.edu.

  5. Where can I get the full data-output from my analysis?

    Select a network, then "Generate/Download Table View". From the Table View, you can select to download the full datatable from the XLinkDB 2.0 database.

Return to top of page.

Protein Modeling and Docking - New to XLinkDB 2.0!


  1. What do I need to do to intiate the modeling/docking on my protein interaction network?

    Nothing! XLinkDB 2.0 has automated the process by which individual structure models and docked structures are generated.

    Once your dataset is uploaded, XLinkDB's job queue will intiate, complete, and report models and docked structures. You can then access these structures within the table view of each dataset.

  2. If no PDB structure exists, what algorithm is used to model and/or dock my protein?

    Currently, XLinkDB uses the Integrative Modeling Platform (IMP) developed by the Sali Lab.

    IMP uses Modeller to generate individual protein structure models and PatchDock to generate docked protein-protein interactions with cross-link distance constraints.

  3. What parameters are used within IMP for modeling and docking?

    For homology modeling, a structural homology model is identified by multiple sequence alignment. This model is then used to predict protein structures.

    For protein-protein docking, cross-link results are used as distance constraints to predict interafacial regions of interaction between multiple proteins.

    We are also currently working to integrate model scoring scripts (e.g. XLMap) into XLinkDB 2.0.

  4. Can I change the parameters used to model and dock my proteins?

    Currently this is not part of XLinkDB 2.0, but we are working towards allowing for this in advanced network inputs.

  5. How can I differentiate empirical protein stuctures (i.e. from the PDB) from modeled protein structures?

    Empirically derived protein structures will be named based on their PDB accession. Modeled protein structures will be named based on their Uniprot accession numbers.

  6. Proteins in my interaction network have empirically derived protein structures in the PDB, but were modeled anyway. Why is this?

    As an example, the PDB protein structure may exist for only part of the protein sequence, but the cross-link data input into XLinkDB exists outside of the protein sequence. Therefore the protein will be submitted for modeling.

Return to top of page.

Network Analysis and Interactomes

  1. What program is used to generate my interaction networks?

    XLinkDB 2.0 uses Cytoscape.js and xiNET to generate protein interaction networks. These graph theory viewers allow for fast and simple analysis of protein-protein interaction networks. - New to XLinkDB 2.0!

  2. Can I filter for a single protein within my dataset within the network view?

    Yes, there are three ways to filter based on: interactors, protein names, and gene label. To filter your network, enter the protein name/Uniprot accession. Once entered a slider will appear. Using the slider, adjust the stringency/leniency of text matching (more or less alike) to whittle down to your protein of interest (e.g. moving the filter bar all the way to the left for a single protein).

  3. Is it possible to see which proteins are interacting with my selected node?

    Yes, within the Cytoscape.js viewer, left-click on your protein node of interest. A small window will appear with all interacting partners for your selected node. Simply click the interactor of interest to explore the exact interactions between these proteins.

  4. How can I save my networks as images?

    From the Cytoscape.js network view, select the link at the bottom of the page to save a PNG image of your network.

  5. I have many links between two proteins, but only one edge is shown, why is this?

    The network is filtered for redundant edges so that only one is shown, though multiple links may exist between or within your proteins of interest. To access your cross-link information either (1) click on "Generate table view" to view your complete dataset or (2) click on the node of interest, then go to the "Properties" tab within the right-hand menu and select the specific interaction you are interested in, once clicked you will be able to see all identified cross-links.

Return to top of page.

Species Level Interactions - New to XLinkDB 2.0!

  1. The full species dataset is slow to load, why is this?

    The full species datasets are cumbersome because they represent multiple thousand interactions from several datasets. The processing time for these will be longer than for a single network. Please be patient in allowing these large networks to load.

  2. How can I determine which datasets a specific cross-link from a species network came from?

    To determine which datasets each cross-link was found in, click on the "Generate table view" button at the top of the network. The table view can be exported, and contains information pertaining to which peptides were found for each protein interaction, and from which dataset each cross-link is derived from.


Return to top of page.