Accurate trajectory inference in time-series spatial transcriptomics with structurally-constrained optimal transport
Bryan, J. P.; Farhi, S. L.; Cleary, B.
Show abstract
New experimental and computational methods use genetic or gene expression observations with single cell resolution to study the relationship between single-cell molecular profiles and developmental trajectories. Most tissues contain spatially contiguous regions that develop as a unit, such as follicles in the ovary, or tubules and glomeruli in the kidney. We find that existing approaches designed to use time series spatial transcriptomics (ST) data produce biologically incoherent trajectories that fail to maintain these structural units over time. We present Spatiotemporal Optimal transport with Contiguous Structures (SOCS), an Optimal Transport-based trajectory inference method for time-series ST that produces trajectory inferences preserving the structural integrity of contiguous biologically meaningful units, along with gene expression similarity and global geometric structure. We show that SOCS produces more plausible trajectory estimates, maintaining the spatial coherence of biological structures across time, enabling more accurate trajectory inference and biological insight than other approaches.
Matching journals
The top 4 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- NiCo Identifies Extrinsic Drivers of Cell State Modulation by Niche Covariation Analysis 97%
- Characterizing cell-type spatial relationships across length scales in spatially resolved omics data 96%
- stLearn: integrating spatial location, tissue morphology and gene expression to find cell types, cell-cell interactions and spatial trajectories within undissociated tissues 96%
Similar papers in this journal
Similar papers in this journal
- Cell region fingerprints enable highly precise single-cell tracking and lineage reconstruction 96%
- Deep generative modeling of transcriptional dynamics for RNA velocity analysis in single cells 96%
- Interpretable representation learning for 3D multi-piece intracellular structures using point clouds 96%
Similar papers in this journal
- scAlign: a tool for alignment, integration and rare cell identification from scRNA-seq data 97%
- geneBasis: an iterative approach for unsupervised selection of targeted gene panels from scRNA-seq. 96%
- Characterizing Spatially Continuous Variations in Tissue Microenvironment through Niche Trajectory Analysis 96%
"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.