Optimal methods for analyzing targeted pairwise knockout screens.

Area of Science:

• Genomics
• Cancer Biology
• Bioinformatics

Background:

• Synthetic lethality targets cancer-specific genetic vulnerabilities for selective tumor cell death.
• Single-gene knockout screens may miss synthetic lethal interactions due to functional redundancy from paralog genes.
• Multiplex CRISPR systems (Cas9, Cas12a) are used for genetic interaction assays, but method comparisons are lacking.

Purpose of the Study:

• To systematically compare bioinformatic approaches for identifying synthetic lethality from CRISPR screens.
• To evaluate the performance of different scoring methods in the context of paralogous genes.
• To identify a robust method for analyzing multiplex CRISPR screening data.

Main Methods:

• Utilized data from four in4mer CRISPR/Cas12a screens in cancer cell lines.
• Applied three bioinformatic approaches: delta log fold change (dLFC), Z-transformed dLFC (ZdLFC), and rescaled dLFC (RdLFC).
• Assessed the consistency of synthetic lethal pair identification across different cell lines for each method.

Main Results:

• Both ZdLFC and RdLFC demonstrated more consistent identification of synthetic lethal pairs compared to the unscaled dLFC method.
• The ZdLFC method provided reproducible results across multiple cancer cell lines.
• No training set of known positive interactors was required for ZdLFC or RdLFC.

Conclusions:

• The ZdLFC method provides a robust framework for scoring synthetic lethal interactions from CRISPR screens involving paralogous genes.
• This method enhances the reliability of synthetic lethality discovery by ensuring consistent results across diverse cell lines.
• ZdLFC offers a valuable tool for cancer research without the need for pre-existing interaction data.