Choose a PPI network
Select the protein-protein interaction source you want to use, such as HPRD, DiamondNet, HINT, or BioGRID.
Tip: Results can differ by network because each source contains different interaction evidence.
User guide
How to use Verdict
Explore relationships among diseases, drugs, groups of proteins, in the protein-protein interaction network.
Workflow
Quick Start
Most analyses follow the same pattern: choose a network, select/build one or more groups, then visualize and interpret the results.
Add groups
Search for diseases, drugs, or proteins. Verdict converts each selection into a protein group.
Example: Add a disease group, then add a drug group to compare them in the interactome.
Visualize
Once your groups are added, click on Visualize to create the network. Use the side panels to inspect proteins, metrics, context, pathways, and enrichment results.
Searching
Search for Diseases, Drugs, or Proteins
Use the tabs on the search page to decide what type of biological entity you want to add.
Diseases
Start typing a disease name and choose a match from the autocomplete list. Verdict adds known disease proteins and the selected number of predicted proteins.
- Use the Quantity of predictions field to control how many predicted proteins are included.
- Predictions appear separately from known or seed proteins in the group table.
Example: Search for "breast cancer" and request 5 predictions.
Drugs
Search for a drug to add its known targets and predicted targets. This is useful for comparing drug action with a disease module or custom protein set.
- Use the prediction quantity to keep the network readable.
- Add several drugs as separate groups when comparing mechanisms.
Example: Search for "canakinumab" and compare it with a disease group.
Proteins
Use the Proteins tab to create a custom group from genes or proteins of interest. You can type multiple proteins or upload a text file with a list of proteins (one per line and represented by its gene symbol).
- Name the group so it is easy to recognize in the network.
- Review any proteins reported as not found before visualizing.
Example: Create a group named "Inflammation markers" with IL1B, IL6, and TNF.
Inputs
Select and Manage Protein Groups
A group can come from a disease, drug, or custom protein list.
Add Group
Adds the selected disease, drug, or protein set to the table. The table shows the group name, color, protein list, controls for including or removing proteins and a delete button for each group.
Update an Existing Group
Use this when you want to add more proteins to a group you already created instead of starting a new group.
Colors
Each group receives a color at random or it can be manually pick by clicking the color picker on the group table. Known or seed proteins use the stronger group color, while predictions use a lighter paired color.
Remove Predictions
Each prediction can be clicked for removing or bringing them back in the table section. Use this to test whether the network interpretation changes when predictions are included.
Visualization
Visualize the Network
The network view shows proteins as nodes and interactions as links. Side panels list group members, network context, metrics, pathways, and enrichment results.
How to read nodes
- Seed genes: known disease genes or known drug targets.
- Predicted genes: predicted genes/targets for the selected disease or drug, or proteins you supplied directly.
- Shared genes: genes that appear in more than one group, often shown with multiple colors.
- Context genes: intermediate genes added to connect or explain relationships in the PPI network.
Screenshot placeholder
Annotated network
Label seed, predicted, shared, and context/intermediate proteins.
Click and Hover
Hover over proteins in the graph or list to highlight the matching node. Click a protein name to open external information such as UniProt or NCBI Gene when available.
Use the side panels
The left panel helps inspect group members and enrichment. The right panel summarizes metrics between groups.
Interpretation
Interpret Scores, Percentiles, Pathways, and ORA
Verdict combines network scores, percentile summaries, pathway membership, and overrepresentation analysis (ORA) to help you reason about biological relationships.
Prediction scores
Prediction scores rank candidate proteins/targets for the selected disease or drug. Higher-ranked proteins are stronger candidates within that search and selected PPI network.
Tip: Compare scores within the same network and input type. Scores from different PPI networks may not be directly interchangeable.
Percentiles
Percentiles show where a score or metric sits relative to a background distribution. For example, P90 means the value is around the 90th percentile of its reference set.
Group and between-group metrics
Metrics such as Commute Time, Diffusion, pStep2, pStep4, and pStep6 summarize network proximity or connectivity patterns within and between groups.
ORA and enrichment results
Overrepresentation analysis asks whether proteins in a group are enriched for known biological categories, such as Gene Ontology terms.
Bars usually represent -log(adjusted p-value), so longer bars indicate smaller adjusted p-values.
Controls
Adjust Visualization Options
Use the controls to make the network easier to inspect, compare, and present.
Layout
Change the graph layout from the top navigation in the visualization view. D3 force is useful for organic network structure; grid and circle layouts can help compare many nodes.
Explore mode
Use Explore when you want to interact with the network more freely, inspect nodes, and follow connections.
Labels
Turn labels on when preparing figures or checking exact gene symbols.
Context level
Context settings control how many intermediate proteins are shown. More context can reveal bridges; less context keeps the display focused on your selected groups.
Colors
Use group colors to keep related proteins visually linked. Choose distinct colors when comparing several diseases or drugs.
Filters and toggles
Use prediction toggles and group controls to ask "does this conclusion depend on predicted proteins, or is it visible using seeds only?"
Export
Download Tables and Network Data
Downloads let you keep a record of your analysis.
Graph Image
Download the current graph view as an image.
All Data
Use this when you need the underlying nodes, links, group assignments, metrics, and related result tables for further analysis.
FAQ
Frequently Asked Questions
Do predicted proteins mean experimentally validated targets?
No. Predicted proteins are computational candidates. They are useful for hypothesis generation and prioritization, but they require biological interpretation and validation.
Why do results change when I choose a different PPI network?
Each PPI network contains different interaction data and coverage. A protein may be highly connected in one network and absent or less connected in another.
How many predictions should I include?
Start small, such as 5 to 10 predictions, then increase if you need broader coverage. Smaller networks are easier to interpret visually.
What should I do with context proteins?
Use them as clues about possible connecting mechanisms. A context protein is not necessarily part of the original disease or drug set, but it may explain how selected proteins are linked.
What does a shared protein indicate?
A shared protein appears in more than one group. It may indicate overlap between disease mechanisms, drug targets, predictions, or custom input sets.
Why is an enrichment result missing?
ORA may be unavailable if too few proteins pass the required thresholds or if no terms are significantly enriched for the current group.
Can I use Verdict results in a publication?
Use the exported tables and network data to document the exact inputs, PPI network, prediction quantity, and interpretation steps. Add methodological details and validation as required for your study.