Co-fluctuations of genes in single cells predict transcriptome-wide outcomes to perturbations

Gene co-fluctuations in unperturbed cells can accurately predict transcriptome-wide responses to perturbations, with CIPHER achieving robust results across large datasets.

• Why it matters: Understanding how genes respond to perturbations is crucial for functional genomics, yet current methods either lack interpretability or ignore gene interactions, limiting insights into cellular responses.
• What they did: CIPHER employs linear response theory to model perturbation outcomes based on baseline gene covariance structures, validated on synthetic networks and applied to over 1.36 million cells across 4,234 perturbations.
• The result: The framework outperforms traditional metrics, reveals conserved fluctuation patterns across experiments, and uncovers fundamental gene modules, enabling more accurate and interpretable predictions of cellular responses.