Skip to main content

Overview

After a virtual screen completes, candidates are available in your experiment for evaluation, triage, and further analysis.

Experiment View

New protein design screen form
The experiment view provides three tabs:
  • Overview - Summary statistics and metrics
  • Table - List view of all candidates with filters and sorting
  • Triage - Organized view for candidate evaluation

Candidate Table

Each candidate row shows:
  • Sequence - Protein sequence with CDR regions highlighted (for antibodies/nanobodies)
  • Id - Unique identifier (e.g., NB-X7CMZA38)
  • Triage - Thumbs down, flag, and thumbs up counts
  • Created By - Which virtual screen generated this candidate
  • Created - Timestamp when the candidate was generated
  • Binding Confidence - Predicted binding affinity score for the target
  • Structure Confidence - Overall structure quality score

Filtering and Selection

  • Default view dropdown - Pre-configured filter sets
  • Columns button - Show/hide table columns
  • Filters button - Add custom filters including:
    • Text filters - Filter by Id, Created By, etc.
    • Number filters - Filter by ipTM, pLDDT, Binding Confidence, Structure Confidence, Loop Fraction using sliders
    • Filter chips show active filters and can be removed individually
  • Select candidates - Use checkboxes to select multiple candidates
  • Download - Export selected candidates

Bulk Actions

When candidates are selected:
  • Add to experiment - Move to another experiment
  • Tag - Apply tags for organization
  • Download - Export candidate data

3D Visualization

The right panel shows:

Target Selection

  • Dropdown to switch between targets
  • Selected candidate IDs displayed as colored tags
  • Download and fullscreen options

3D Structure Viewer

  • Interactive 3D ribbon diagram showing:
    • Target (light grey)
    • Binder candidates (colored by selection)
  • 3D axis indicator for orientation
  • Rotate, zoom, and pan controls

Properties Comparison

Compare metrics across selected candidates:
  • ipTM - Interface predicted Template Modeling score
  • pLDDT - Predicted Local Distance Difference Test (structure confidence)
  • Binding Confidence - Predicted binding affinity
  • Structure Confidence - Overall structure quality
Each chart shows all selected candidates as colored lines, making it easy to identify top performers.

Triage View

The Triage tab provides a detailed view for evaluating individual candidates:

Left Panel - Candidate Details

  • Sequence Viewer:
    • Full protein sequence with Framework Regions (FR) and Complementarity Determining Regions (CDR) highlighted
    • CDR regions color-coded (CDR1, CDR2, CDR3)
    • Residue numbers displayed above sequence
    • Navigation arrows to move between candidates
  • Candidate Actions:
    • Star - Favorite a candidate
    • Thumbs down - Reject candidate
    • Flag - Mark for further review
    • Thumbs up - Approve candidate
    • + Tag - Apply tags for organization
  • Properties and Descriptors:
    • Properties tab - View metrics (ipTM, pLDDT, Binding Confidence, Structure Confidence)
    • Descriptors tab - View structural features (Loop Fraction, Helix Fraction, Sheet Fraction)
    • Switch between Table and Plot views for each

Right Panel - 3D Structure and Comparison

  • 3D Structure Viewer:
    • Shows the selected candidate’s binding pose with the target
    • Binder colored orange, target colored light grey/green
    • Interactive controls for rotation, zoom, and pan
  • Properties Comparison:
    • Parallel coordinates plot showing all candidates
    • Selected candidate highlighted in purple
    • Non-selected candidates shown in dark grey
    • Compare across multiple properties simultaneously

Triage Workflow

  1. Select candidates - Use checkboxes or filters to narrow down
  2. Review 3D structures - Examine binding poses in the viewer
  3. Compare properties - Use charts to identify high-scoring candidates
  4. Triage actions:
    • Thumbs down - Reject candidate
    • Flag - Mark for further review
    • Thumbs up - Approve candidate
  5. Tag candidates - Apply tags for organization (e.g., “high-affinity”, “interesting-pose”)
  6. Export - Download selected candidates for further analysis

Tips

  • Start by filtering to top performers (high binding confidence)
  • Review 3D structures to ensure poses are reasonable
  • Compare multiple candidates side-by-side using the properties charts
  • Use tags to organize candidates by different criteria
  • Export promising candidates for experimental validation