SynLeGG

SynLeGG identifies genetic dependency relationships, particularly synthetic lethality, by integrating transcriptome, CRISPR, genomic, proteomic, protein–protein interaction, evolutionary and Gene Ontology data across 783 cancer cell lines and 30 tissues to prioritize therapeutic targets for precision oncology.

Key Features:

• MultiSEp algorithm: MultiSEp performs unsupervised clustering of cell lines based on gene expression profiles.
• Data integration: Integrates CRISPR screens, gene expression, genomic information, proteomics, protein–protein interactions, evolutionary data, and Gene Ontology annotations for discovery and interpretation of genetic dependency relationships.
• Transcriptome-based discovery: Leverages transcriptome data to propose candidate genetic dependencies across 783 cancer cell lines and 30 tissues.
• CRISPR and mutation analysis: Combines MultiSEp-derived clusters with CRISPR scores and mutational status to identify potential genetic dependency relationships.
• Pan-cancer and tissue-specific analysis: Supports both pan-cancer analyses and tissue-specific GDR discovery across 30 tissues.
• Benchmarking against SynLethDB: Demonstrates higher area under the Receiver Operator Characteristic (ROC) curve and 2.8- to 8.5-fold improved coverage compared to SynLethDB.
• Recovery of known relationships: Recovers known synthetic lethality pairs such as SMARCA2–SMARCA4.
• Proteomics and paralogue enrichment: Enriches predictions with proteomics data and paralogue information to aid prioritization of candidates.
• CRISPR validation enrichment: Predicted dependencies are significantly enriched for validation by recent CRISPR screens.

Scientific Applications:

• Therapeutic target prioritization: Prioritization of candidate therapeutic targets and synthetic lethal pairs for precision oncology research.
• Genetic vulnerability identification: Identification of genetic vulnerabilities and genetic dependency relationships across cancer cell lines and tissues.
• Dependency interpretation: Interpretation and prioritization of candidate dependencies using Gene Ontology annotations, protein–protein interaction networks, proteomics and paralogue information.
• Tissue-specific discovery: Discovery of tissue-specific genetic dependencies across 30 tissue types.
• Benchmarking and validation: Benchmarking of predicted dependencies against SynLethDB and validation against CRISPR screen results.

Methodology:

SynLeGG uses the MultiSEp unsupervised algorithm to cluster cell lines by gene expression; these clusters are analyzed with CRISPR scores and mutational status to identify genetic dependency relationships, and predictions are benchmarked against SynLethDB using area under the ROC curve and coverage comparisons while integrating proteomics, protein–protein interaction, evolutionary and Gene Ontology annotations for enrichment and interpretation.