Search papers, labs, and topics across Lattice.
The paper introduces AROMA, a multimodal architecture for virtual cell genetic perturbation modeling that integrates textual evidence, graph-topology information, and protein sequence features. AROMA is trained using a two-stage optimization strategy to predict molecular state changes under genetic perturbations, improving both accuracy and interpretability. Experiments demonstrate that AROMA outperforms existing methods across multiple cell lines, exhibits robustness in zero-shot settings, and handles knowledge-sparse scenarios effectively.
Achieve more reliable and interpretable virtual cell perturbation predictions by combining knowledge-driven multimodal modeling with evidence retrieval.
Virtual cell modeling predicts molecular state changes under genetic perturbations in silico, which is essential for biological mechanism studies. However, existing approaches suffer from unconstrained reasoning, uninterpretable predictions, and retrieval signals that are weakly aligned with regulatory topology. To address these limitations, we propose AROMA, an Augmented Reasoning Over a Multimodal Architecture for virtual cell genetic perturbation modeling. AROMA integrates textual evidence, graph-topology information, and protein sequence features to model perturbation-target dependencies, and is trained with a two-stage optimization strategy to yield predictions that are both accurate and interpretable. We also construct two knowledge graphs and a perturbation reasoning dataset, PerturbReason, containing more than 498k samples, as reusable resources for the virtual cell domain. Experiments show that AROMA outperforms existing methods across multiple cell lines, and remains robust under zero-shot evaluation on an unseen cell line, as well as in knowledge-sparse, long-tail scenarios. Overall, AROMA demonstrates that combining knowledge-driven multimodal modeling with evidence retrieval provides a promising pathway toward more reliable and interpretable virtual cell perturbation prediction. Model weights are available at https://huggingface.co/blazerye/AROMA. Code is available at https://github.com/blazerye/AROMA.
