#37 · Single-cell

Inferring cell trajectories and tissue heterogeneity from single-cell transcriptomes

The dominant thread is computational inference of cell-state trajectories and heterogeneity from single-cell transcriptomic data. TRIO-AI integrates temporal graph neural networks, neural ODEs, and time-VAEs to reconstruct macrophage trajectories in mouse liver ischemia-reperfusion injury from scRNA-seq. The carotid atherosclerosis study combines single-cell and spatial transcriptomics with new segmentation algorithms to map morphological heterogeneity and macrophage transdifferentiation within plaque. What is new here is the coupling of trajectory inference to temporal dynamics and to spatial tissue context in macrophage-centric injury and plaque systems. A multiscale hemodynamic model of cerebral circulation is also among the items, but it is not single-cell and does not share this thread.