![[1] Predicting cell morphological responses to perturbations using generative modeling (IMPA): Nature Communications (2025) https://www.nature.com/articles/s41467-024-55707-8
IMPA employs a conditional style transfer process to separate perturbation and batch information from basic cellular information and incorporates perturbation details. It demonstrated significant improvement over other batch correction methods on the RxRx1 dataset and showed the ability to generalize to unseen perturbations.](https://pbs.twimg.com/media/GnX_RVKW0AAPWjC.jpg?name=small&format=webp)
ALT [1] Predicting cell morphological responses to perturbations using generative modeling (IMPA): Nature Communications (2025) https://www.nature.com/articles/s41467-024-55707-8 IMPA employs a conditional style transfer process to separate perturbation and batch information from basic cellular information and incorporates perturbation details. It demonstrated significant improvement over other batch correction methods on the RxRx1 dataset and showed the ability to generalize to unseen perturbations.
![[2] MorphoDiff: Cellular Morphology Painting with Diffusion Models: ICLR2025 (accepted spotlight presentation) https://openreview.net/forum?id=PstM8YfhvI
MorphoDiff utilizes a diffusion process to generate images while encoding perturbation information in the latent space. It excelled in generating images with high fidelity and demonstrated the ability to generalize to held-out compounds.](https://pbs.twimg.com/media/GnX_o9zXYAAFysp.jpg?name=small&format=webp)
ALT [2] MorphoDiff: Cellular Morphology Painting with Diffusion Models: ICLR2025 (accepted spotlight presentation) https://openreview.net/forum?id=PstM8YfhvI MorphoDiff utilizes a diffusion process to generate images while encoding perturbation information in the latent space. It excelled in generating images with high fidelity and demonstrated the ability to generalize to held-out compounds.

















