CellStream: Dynamical Optimal Transport Informed Embeddings for Reconstructing Cellular Trajectories from Snapshots Data

2025-11-16

AI Summary

Static single-cell RNA sequencing (scRNA-seq) snapshots poorly capture continuous transcriptional dynamics due to data sparsity, high technical noise, and the common practice of decoupling trajectory inference from embedding learning—thereby neglecting inherent temporal structure. Method: We propose DyOT, an end-to-end deep generative model that jointly learns dynamic-aware low-dimensional embeddings and cellular evolutionary trajectories by integrating unbalanced optimal transport (OT) with an autoencoder architecture. Its core innovation lies in modeling developmental time as an OT flow, explicitly enforcing continuous dynamical constraints in the latent space. Results: On multiple simulated and real developmental datasets—including pseudotime and organogenesis benchmarks—DyOT achieves significantly improved temporal coherence and robustness of inferred trajectories. It outperforms state-of-the-art methods (e.g., Slingshot, PAGA, VeloAE) in noise tolerance and provides an interpretable, generalizable framework for dissecting single-cell dynamics.

Technology Category

• Machine Learning: Deep Generative Models & Autoencoders
• Search and Optimization: Learning to Search
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Web Mining and Content Analysis: Models for Web evolution
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Single-cell RNA sequencing (scRNA-seq), especially temporally resolved datasets, enables genome-wide profiling of gene expression dynamics at single-cell resolution across discrete time points. However, current technologies provide only sparse, static snapshots of cell states and are inherently influenced by technical noise, complicating the inference and representation of continuous transcriptional dynamics. Although embedding methods can reduce dimensionality and mitigate technical noise, the majority of existing approaches typically treat trajectory inference separately from embedding construction, often neglecting temporal structure. To address this challenge, here we introduce CellStream, a novel deep learning framework that jointly learns embedding and cellular dynamics from single-cell snapshot data by integrating an autoencoder with unbalanced dynamical optimal transport. Compared to existing methods, CellStream generates dynamics-informed embeddings that robustly capture temporal developmental processes while maintaining high consistency with the underlying data manifold. We demonstrate CellStream's effectiveness on both simulated datasets and real scRNA-seq data, including spatial transcriptomics. Our experiments indicate significant quantitative improvements over state-of-the-art methods in representing cellular trajectories with enhanced temporal coherence and reduced noise sensitivity. Overall, CellStream provides a new tool for learning and representing continuous streams from the noisy, static snapshots of single-cell gene expression.

Problem

• Reconstructing continuous cellular trajectories from sparse static snapshots
• Integrating trajectory inference with embedding construction using optimal transport
• Reducing technical noise while maintaining temporal coherence in embeddings

Innovation

• Integrates autoencoder with unbalanced dynamical optimal transport
• Learns embedding and cellular dynamics jointly from snapshots
• Generates dynamics-informed embeddings capturing temporal developmental processes