Chronos

Chronos infers gene fitness effects from CRISPR knockout experiments by explicitly modeling cell proliferation dynamics to improve estimation of gene-specific fitness over time.

Key Features:

• Explicit Modeling of Cell Proliferation Dynamics: Chronos models temporal cell proliferation dynamics following CRISPR gene knockouts to provide more accurate inference of gene fitness effects.
• Utilization of Longitudinal Data: Chronos leverages longitudinal CRISPR read count data, performing competitively with read counts from a single late time point and improving inference with multiple late time points.
• Bias Mitigation: Chronos accounts for confounding factors including DNA cutting toxicity, incomplete phenotype penetrance, and screen quality bias.
• Cross-Screen Information Sharing: Chronos shares information across screens for joint analysis, enabling integration of large datasets such as Project Achilles and Score.
• Empirical Performance: Chronos has been demonstrated to outperform competing methods across various performance metrics in different experiment types.

Scientific Applications:

• Cancer Biology: Analysis of CRISPR loss of function screens to identify gene dependencies and interactions in cancer.
• Large-Scale Screen Integration: Joint analysis and integration of large CRISPR screening projects and datasets such as Project Achilles and Score to enhance biological insight.

Methodology:

Chronos models the temporal dynamics of cell proliferation after CRISPR knockout and integrates longitudinal read counts and cross-screen information to mitigate biases such as DNA cutting toxicity, incomplete phenotype penetrance, and screen quality bias.