FlyPredictome Network

Interactive network viewer for the Drosophila protein-protein interaction atlas.

Open the viewer

Overview

This viewer enables interactive exploration of an evidence-supported PPI network from the FlyPredictome project. The network contains 4,551 proteins and 18,728 interactions organized into 29 clusters with hierarchical sub-clustering.

Network construction

1. Drosophila protein pairs were scored with AlphaFold-Multimer and iLIS
2. Positive PPIs were selected at iLIS >= 0.223 (10% FDR)
3. Only PPIs with prior evidence were retained: fly literature, BioGRID, genetic interaction (MIST), or conserved interolog (DIOPT >= 4)
4. Proteins with >= 2 connections were kept (degree >= 2 filter)
5. The giant connected component was extracted
6. Louvain community detection identified clusters, sub-clusters, and sub-sub-clusters

Features

• 3-level drill-down: Cluster -> Sub-cluster -> Sub-sub-cluster
• Cluster-aware layout: Sub-clusters positioned by inter-group connectivity
• Evidence filter: All / Fly only / Interolog only / Both
• FDR filter: <= 10% / <= 5% / <= 1%
• Edge coloring: Blue (fly evidence), Teal (both), Orange (interolog)
• Gene search: Search by gene symbol or FBgn ID
• Multi-gene analysis: Select genes -> Find Paths / Shared neighbors
• iLIS scores: Shown for each interaction
• External links: FlyBase, FlyPredictome database
• Export: PNG (white/transparent background) and SVG
• Adjustable display: Node size, font size, edge opacity/width sliders

Files

File	Description
index.html	Self-contained viewer (just open in browser)
viewer_template.html	HTML/CSS/JS template
generate_html.py	Embeds data into template to produce index.html
data/nodes.csv	Gene nodes with cluster/sub-cluster assignments
data/edges.csv	PPI edges with iLIS scores, evidence type, FDR
data/clusters.json	Cluster metadata (3-level hierarchy)
data/layouts.json	Pre-computed layout coordinates

Rebuilding

To regenerate index.html after modifying data or template:

python3 generate_html.py

Requires Python 3 with no external dependencies.

Data sources

• FlyPredictome database: https://www.flyrnai.org/tools/fly_predictome
• iLIS scoring pipeline: https://github.com/flyark/AFM-LIS

Citation

Kim et al., FlyPredictome: a structurally resolved Drosophila interactome atlas. (2026)

License

MIT License