PerturbReason: Training Data for a Multimodal Virtual Cell Genetic Perturbation Model

PerturbReason is the training dataset for the AROMA model, a multimodal architecture for virtual cell modeling presented at ACL 2026. The dataset integrates textual evidence, graph topology, and protein sequences to predict the effects of genetic perturbations. It was authored by blazerye and last updated on Hugging Face in April 2026.

Use Cases

• Training multimodal AI models for genetic perturbation prediction based on integrated textual, graph, and sequence data.
• Benchmarking virtual cell modeling architectures based on the described multimodal evidence sources.
• Researching the integration of knowledge graphs and protein sequences for biological AI systems.

Strengths

• Dataset is associated with a peer-reviewed publication (ACL 2026 Findings).
• Data is designed for a novel multimodal architecture integrating multiple evidence types.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

blazerye on Hugging Face, associated with the AROMA model research.

Collection Method

Likely created as training data for the described AROMA model.

Time Range

null

Freshness

Last updated 2026-04 23 06:18:45; freshness should be verified.

Geography

null