Counterfactual Hypothesis Testing of Tumor Microenvironment Scenarios Through Semantic Image Synthesis
Li, D.; Ma, Q.; Chen, J. L.; Liu, A.; Cheung, J.; Xie, Y.; Gudjonson, H.; Nawy, T.; Pe'er, D.; Pe'er, I.
Show abstract
AO_SCPLOWBSTRACTC_SCPLOWRecent multiplexed protein imaging technologies characterize cells, their spatial organization, and interactions within microenvironments at an unprecedented resolution. Although observational data can reveal spatial associations, it does not allow users to infer salient biological relationships and cellular interactions. To address this challenge, we develop a generative model that allows users to test hypotheses about the effect of cell-cell interactions on protein expression through in silico perturbation. Our Cell-Cell Interaction GAN (CCIGAN) model employs a generative adversarial network (GAN) architecture to generate high fidelity synthetic multiplexed images from semantic cell segmentations. Our approach is unique in that it learns relationships between all imaging channels simultaneously and yields biological insights from multiple imaging technologies in silico, capturing known tumor-immune cell interactions missed by other state-of-the-art GAN models.
Matching journals
The top 5 journals account for 50% of the predicted probability mass.
Similar papers in this journal
Similar papers in this journal
- Learning tissue representation by identification of persistent local patterns in spatial omics data 96%
- Compressed sensing expands the multiplexity of imaging mass cytometry 96%
- scDREAMER: atlas-level integration of single-cell datasets using deep generative model paired with adversarial classifier 96%
Similar papers in this journal
- S3-CIMA: Supervised spatial single-cell image analysis for the identification of disease-associated cell type compositions in tissue 96%
- CellContrast: Reconstructing Spatial Relationships in Single-Cell RNA Sequencing Data via Deep Contrastive Learning 95%
- Bi-level Graph Learning Unveils Prognosis-Relevant Tumor Microenvironment Patterns in Breast Multiplexed Digital Pathology 95%
Similar papers in this journal
Similar papers in this journal
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.