RosettaSX

RosettaSX scores gene expression signatures in cancer models and patient datasets to quantify oncogenic signaling pathway activities, infer cellular origins of tumors, and assess immune cell infiltration.

Key Features:

• Signature scoring: Scores gene expression signatures at the individual tumor/sample level across cancer models and patient cohorts.
• Coherent/translatable signature identification: Identifies coherent gene expression within signatures and selects translatable signatures before scoring to reassess applicability in new experimental contexts.
• Hand-curated gene sets: Leverages 293 hand-curated literature-derived gene sets that span transcriptional modules relevant to cancer.
• Reference data integration: Leverages expression data from the Cancer Cell Line Encyclopaedia (CCLE) and The Cancer Genome Atlas (TCGA) for scoring and comparison.
• Molecular data augmentation: Integrates single nucleotide variants (SNVs), copy number variations (CNVs), gene expression levels, gene dependency data, and protein abundance alongside signature scores.
• Visualization outputs: Produces per-cancer signature heat maps, clustered heatmaps, and additional plotting outputs for comparative and exploratory analysis.
• Cross-cancer and subtype analysis: Enables comparative exploration of signature activity across cancer subtypes and between cancer types.

Scientific Applications:

• Oncogenic pathway activity profiling: Quantifies activities of oncogenic signaling pathways from expression signatures.
• Cellular origin inference: Infers cellular origins of tumors using transcriptional modules.
• Immune infiltration assessment: Assesses immune cell infiltration into tumor tissues via immune-related expression signatures.
• Genomic–transcriptomic integration: Associates SNVs, CNVs, gene dependency, and protein abundance with signature activity to explore mechanistic links.
• Comparative oncology studies: Supports cross-cancer and subtype comparisons to investigate interactions among oncological mechanisms.

Methodology:

Identifies coherent, translatable gene expression signatures from large signature collections and then scores their activity in individual tumors using those selected signatures.