Learning how single cells respond to perturbations.

State, a machine learning model, predicts single-cell transcriptome responses to perturbations across diverse cellular contexts.

• Why it matters: Understanding how individual cells respond to genetic or chemical changes is crucial for advancing cell biology and therapeutic strategies. However, predicting these responses across different cell types remains challenging due to cellular heterogeneity.
• What they did: Adduri et al. developed State, a machine learning approach that generalizes across various cellular contexts to forecast transcriptomic responses to perturbations, demonstrating its broad applicability.
• The result: State enables more accurate predictions of single-cell responses, facilitating improved understanding of cellular behavior and potentially guiding targeted interventions in complex biological systems.