TARGET-SL: precision essential gene prediction using driver prioritisation and synthetic lethality.

Area of Science:

• Bioinformatics
• Cancer Genomics
• Synthetic Lethality

Background:

• Identifying patient-specific vulnerabilities is crucial for targeted cancer therapies.
• Current predictive bioinformatics tools lack robust validation, hindering clinical translation.
• Personalized driver prioritization algorithms (PDPAs) provide patient-specific insights but are difficult to translate into actionable treatment strategies.

Purpose of the Study:

• To develop a framework for benchmarking and validating personalized driver prioritization algorithms (PDPAs).
• To introduce the Tumour-specific Algorithm for Ranking GEnetic Targets via Synthetic Lethality (TARGET-SL) for identifying and validating patient-specific essential genes.
• To improve the clinical translatability of predictive bioinformatics tools in oncology.

Main Methods:

• Developed the TARGET-SL framework utilizing PDPA predictions to generate ranked lists of essential genes.
• Implemented a novel benchmarking strategy comparing PDPA predictions against ground truth gene essentiality data.
• Utilized large-scale CRISPR-knockout and drug sensitivity screening data for validation.

Main Results:

• TARGET-SL effectively benchmarks PDPAs by comparing predictions with experimental essentiality data.
• The framework identifies tumor-exclusive vulnerabilities, outperforming predictions based on canonical driver genes.
• TARGET-SL demonstrates superior performance in identifying sample-specific vulnerabilities compared to existing tools.

Conclusions:

• TARGET-SL provides a validated approach to identify and prioritize patient-specific cancer vulnerabilities.
• This framework enhances the clinical utility of predictive bioinformatics tools for personalized cancer treatment.
• TARGET-SL facilitates the translation of genomic insights into actionable therapeutic strategies through robust in vitro and in vivo validation.