• Displays global counts: total cancers, drugs, drug groups, and samples.
• Provides a global search bar (e.g., cancer names or sample accessions such as GEO IDs).
• Offers quick navigation to Cancer, Drug Group, and scRNA-seq dataset pages. Search is case-insensitive and supports partial matches.

Each cancer page summarizes:

• Overview cards: tissue, number of samples, drugs, and drug groups. Higher sample counts support more stable estimates of pathway activity and drug associations; broader drug/drug-group coverage expands the hypothesis space for repurposing.
• Drug groups: cards list representative drugs and hub-gene tags per group. Representative drugs summarize a group’s mechanism of action (MoA) at a glance. Hub genes highlight putative regulatory bottlenecks in the group’s molecular context, guiding target-focused follow-up.
• Drug cluster heatmap: pathway-level clustering across drugs. Clustering by pathway signatures reveals shared biological effects and potential MoA convergence.
• Cancer-level pathway bubble chart: bubble size (−log₁₀ p) and color (NES) prioritize pathways central to the cancer context. This view supports hypothesis generation about processes to perturb (e.g., immune evasion, metabolic rewiring) and aligns drugs with those processes.
• Gene Regulatory Network (GRN): TFs and targets are rendered in distinct colors; edges can be filtered by importance and top-K per TF. A CSV download is available for offline analysis. Mapping transcription factors to their targets exposes regulatory modules underlying observed pathway signals. High-importance edges and TF hubs suggest leverage points where pharmacologic or genetic perturbation may yield maximal downstream effect.

For individual single-cell datasets, you can:

• Filter by cancer type, tissue, and/or sample ID.
• Open a sample to view its metadata (e.g., instrument model, cell counts, tissue).
• Inspect candidate comparison samples (same cancer) suggested inline. Proportional shifts in immune, stromal, and malignant compartments illuminate tumor-microenvironment heterogeneity, which often modulates drug response and can explain divergent connectivity scores across samples.
• View the drug connectivity score table. Scores, p-values, and FDR provide a ranked, multiple-testing-aware list of candidate compounds for follow-up. Combining ranks with cell-type context helps prioritize agents most likely to act on the dominant biology of the sample.

• Enrichment bubble chart (Drug × Pathway): x = drug, y = pathway description, color = NES (diverging scale), size = −log₁₀(p). Includes an ontology filter (GO: BP/CC/MF) and letter-window sliders for paging axes. This links individual drugs to functional pathways and clarifies whether compounds converge on common biology.
• PPI network: multiple edge layers reflect evidence types (e.g., neighborhood, fusion, co-occurrence, co-expression, experimental, database, text-mining). Opacity and width scale with the evidence score; COSE/Grid/Circle layouts are available, and the legend indicates layer meanings. Visualizing proteins and their multi-evidence interactions highlights modules that may mediate drug effects. Highly connected subgraphs suggest core processes; layering evidence increases confidence in key edges.
• cytoHubba ranking: sortable table with centrality metrics (MCC, DMNC, MNC, Degree, EPC, Bottleneck, Eccentricity, Closeness, Radiality, Betweenness, Stress, Clustering Coefficient). Supports search, pagination, and ranking. Centrality metrics prioritize putative key regulators for experimental validation. Use with enrichment and GRN views to mitigate degree bias and select targets supported across modalities.

• Upload single-cell expression matrices in one of the supported formats: CSV, TSV, RDS (Seurat), H5/H5Seurat, or 10x .tar.gz.
• Provide required metadata: cancer type and tissue. Please supply an email address to receive the results.
• Pipeline stages: uploading → preprocess → annotation → drug_repurposing → done (or failed on error).
• Email delivery includes the key outputs (e.g., NORM RDS, annotation results, drug-scoring results).