The Bioverse Viewer may be accessed through the Bioverse framework, where you can visualise predicted protein interaction networks for more than 50 organisms with detailed annotations, and through a dedicated Bioverse Viewer interface, where you also have the ability to upload user-supplied interactome data.
A typical graph view consists of a toolbar and a graph window. The toolbar contains a search field, freeze option (which stops the relaxation layout algorithm), drop-down functional annotation menu, distance selector (distance from selected node), screen refresh button, and total node/edge count. Test each feature out to see what it does.
Searches can be performed by typing in the exact node label you're looking for, or by using a wildcard asterisk (*) character. Note that wildcarding can only be used following a specific character, and only once per search term. If the search fails, a random set of nodes will appear. If no matches are found, a message indicating "No match" will appear in the search field. Search terms are case-sensitive.
Using the left mouse button, select a specific node. It will turn green. All neighboring nodes up to a distance selected will highlight as yellow. Nodes outside of this neighborhood will be gray. Undirected edges with weights will appear according to a color code where green denotes the highest confidence, red denotes the lower, and orange and yellow denote intermediate confidences.
Individual nodes can also be dragged by holding down on the left button on a specific node. Double-clicking on a node will isolate the immediate neighbors around that node, removing other nodes that lie outside of the neighborhood set by the distance menu. Right-button clicking on a node will display a hyperlinked record file for a given node, as well as expand that node for neighbors. The URL path, as well as all the relevant node and edge data, is stored as HTML parameters in the page that displays the applet.
A drop-down menu to select only proteins in particular functional category is available. Graph navigation can begin as before.
The dedicated
Bioverse Viewer interface enables you to upload user-supplied interactome
data and visualise it using our interface. You can upload a list of
interactions and confidences in a comma-separated value (csv) file
format. You can also optionally upload a list descriptions for each of
the nodes in the interaction file. Detailed description of these
formats is given in the documentation for the
csv2html
script, and example
interaction and
function list files in csv
format are available.
The source code may be downloaded at http://software.compbio.washington.edu/bioverse_viewer/bioverse_viewer.tgz. An HTML file listing the details of the interactome needs to be created and the viewer called from within the HTML file. Be sure to modify the codebase parameter properly to the path containing the compiled ViewerApplet.class file.
HTML pages should embed the node, edge, and functional data using
the following format. Note that the three key data parameters are the
functionlist
, interaction
, and
functionkey
. These parameters are consistent with the
original Mrowka viewer. However, please note that the contents of the data
are different in our version. functionlist
relate node labels
to molecule identifiers. In the
Bioverse framework, we employ a numerical identifier system for proteins. The
interaction
parameter contains the edges between nodes, using
the node labels, separated by the edge weights. Finally, the
functionkey
contains two pieces of data. The first parameter,
if provided, is the URL pattern used to hyperlink individual nodes to
other detailed record pages. The rest of the data contains the
functional classes that each protein (identified by record ID) belongs
to. This data is used for the functional annotation drop down
menu.
An example HTML page that is commented in detail is available at http://bioverse.compbio.washington.edu/viewer/example.html. Please review the comments in the code by viewing the source. Note that this is only a fragment of a file, used for instructional purposes. To see an actual working file, look at the page source from a Bioverse predicted network. It is also important to keep in mind that this viewer can at most display 400 nodes simultaneously.
Usage: csv2html
interaction-file
[
function-list-file]
Rather than try to figure out how to code the HTML files, this
convenient program, available in the src/scripts
directory in
the
source distribution, can be used. It takes two comma-separated
value (csv) files as input and generates an HTML page that can be used
with the Bioverse Viewer. The first file specifies the list of
interactions in this form:
molecule1,molecule2,confidence
molecule2,molecule3,confidence
molecule3,molecule1,confidence
where confidence is a value ranging from 0 to 1 indicating the strength of the relationship between the two molecules denoted by a single-word identifier.
The second file, which is optional, is a list of descriptions (such as functional descriptions) for the interacting molecules from the first file. This is specified in this form:
molecule1,description1
molecule2,description2
molecule3,description3
where description may be an arbitrary string containing any of the 26 letters in the English alphabet. Example interaction and function list files in csv format are available.