Multi-domain translation between single-cell imaging and sequencing data using autoencoders

The development of single-cell methods for capturing different data modalities including imaging and sequencing has revolutionized our ability to identify heterogeneous cell states. Different data modalities provide different perspectives on a population of cells, and their integration is critical for studying cellular heterogeneity and its function. While various methods have been proposed to integrate different sequencing data modalities, coupling imaging and sequencing has been an open challenge. We here present an approach for integrating vastly different modalities by learning a probabilistic coupling between the different data modalities using autoencoders to map to a shared latent space. We validate this approach by integrating single-cell RNA-seq and chromatin images to identify distinct subpopulations of human naive CD4+ T-cells that are poised for activation. Collectively, our approach provides a framework to integrate and translate between data modalities that cannot yet be measured within the same cell for diverse applications in biomedical discovery. Integration of single cell data modalities increases the richness of information about the heterogeneity of cell states, but integration of imaging and transcriptomics is an open challenge. Here the authors use autoencoders to learn a probabilistic coupling and map these modalities to a shared latent space.

Automated summary

The authors use autoencoders to learn a probabilistic coupling and map different data modalities to a shared latent space, presenting an approach for integrating vastly different modalities. The integration of imaging and transcriptomics is still an open challenge, but this method provides a framework for diverse applications in biomedical discovery.