User Guide

  1. Overview

    GPS-Prot searches databases and displays results in the Navigator window.

    Double click any protein in the network to expand the network.

    Increase or decrease the number of proteins shown with the Confidence Score slider. Confidence scores are calculated as described below.

    Color Key: PPIs are colored Green. Proteomic/genomic screen hits (implemented for HIV only at this time) can be overlaid on the existing PPI network. Such nodes are colored Yellow. Create an account to upload your own Proteomic/Genomic screen hits or PPIs.

  2. The Navigator window

    Graph Panel: Adjust areas by clicking and dragging borders. Double click any protein to search databases and display binding partners. Single click to see information. “Clusters” is an option that highlights clusters of proteins in which each protein binds to all the others.*

    Upper Left Panel: Displays information about a Protein or Interaction. Each interaction comes with a link to publications on PubMed. Access the entire publication history for an interaction with one click.

    Lower Left Table: Lists information for Proteins or Interactions in the network. Columns in table can be sorted, for example to sort by score, or protein identifier. The Search box searches the table for proteins in the presently-loaded network. It does NOT launch a new network. To launch a new network, enter gene identifier in the left panel.

    *IMPORTANT!!! What you see in Navigator is not the complete PPI Network. It contains only the proteins that bind to your initial search protein or nodes you expanded during your session. You must double click on a protein to be sure you are seeing ALL of its partners. This may affect assignment to clusters.

  3. Methods

    1. Computational Methods: Data retrieved from external databases resides in a MySQL relational database. Identifiers are mapped to GeneName/UniprotID pairs. This allows us to distinguish between multiple proteins with the same name (e.g. HLA-A) and also to distinguish between sub-proteins that share the same Uniprot ID (e.g. PR, RT, IN). The logic tier is handled by PHP5 and the output of each database search is JSON data describing individual proteins and binary interactions. The JSON data is rendered in the viewer using D3.

    2. Duplicate Experiments: Where duplicates exist in databases, they are consolidated and scored once (see below).

    3. Confidence Scores: High quality confidence scores for human-human interactions are retrieved from the HIPPIE database. Crude confidence scores for HIV-human interactions are calculated based on the number of publications for each interaction in the HHPID database, currently the most comprehensive and frequently updated database for HIV-host interactions.

      1. Human-human interaction scoring: See HIPPIE database (Schaefer et al., 2012, PLoS One, 7: e31826) for details of human-human interaction scoring.
      2. HIV-human interaction scoring: The score, S, (for HIV-host interactions) is calculated as S = 1 - a-x, where a is 1.3 and x is the sum of the number of unique publications reporting the interaction.
      3. CORUM data and user-uploaded data are scored as 1.0 so they are visible in all networks, regardless of filtering.