pith. sign in

arxiv: 2605.22340 · v1 · pith:5JB2GMXKnew · submitted 2026-05-21 · 💻 cs.LG

From Snapshots to Trajectories: Learning Single-Cell Gene Expression Dynamics via Conditional Flow Matching

Siyu Pu , Qingqing Long , Xiaohan Huang , Haotian Chen , Jiajia Wang , Meng Xiao , Xiao Luo , Hengshu Zhu
show 2 more authors
Yuanchun Zhou Xuezhi Wang
This is my paper

Pith reviewed 2026-05-22 07:08 UTC · model grok-4.3

classification 💻 cs.LG
keywords single-cell RNA-seqflow matchingtrajectory inferenceoptimal transportgenerative modelingcellular dynamicstime-series data
0
0 comments X

The pith

A conditional flow matching model learns continuous gene expression dynamics from discrete unpaired cell snapshots.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

This paper tries to show that gene expression trajectories can be learned directly from time-series scRNA-seq data collected as separate snapshots at different times. It does so by creating a flow matching model that uses optimal transport to softly match cells between time points and learns how the distribution evolves. If successful, this would let researchers fill in the missing time points and visualize or predict how cells change without dense measurements, which is useful for studying processes like cell differentiation or tumor evolution where only a few time points are feasible to collect.

Core claim

The central claim is that coupling-conditioned flow matching, by constructing soft targets from entropically regularized optimal transport couplings between adjacent snapshots, learning bidirectional velocity fields with consistency refinement, and applying distribution alignment and regularization, enables accurate temporal interpolation, extrapolation, and trajectory reconstruction while mitigating distribution drift in long-horizon predictions.

What carries the argument

coupling-conditioned flow matching that uses entropically regularized OT couplings as soft targets for time-dependent velocity fields

If this is right

  • Consistent improvements in distributional prediction accuracy for both interpolating between observed time points and extrapolating beyond them.
  • More precise reconstruction of individual cell trajectories even when many intermediate time points are missing.
  • Temporally coherent visualizations of how gene expression changes over time.
  • Reduced drift in predicted distributions during extended rollouts of the dynamics.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same conditioning strategy might help in modeling dynamics from other types of snapshot data, such as in ecology or social networks.
  • Combining this with spatial information could lead to models of tissue-level dynamics.
  • Validating against ground-truth trajectories from live-cell imaging would test the fidelity of the inferred dynamics.

Load-bearing premise

Entropically regularized optimal transport couplings between snapshots can serve as reliable soft targets for velocity field learning, and the added consistency and alignment steps are enough to stop error accumulation over long time periods.

What would settle it

If one generates synthetic data with known continuous dynamics, samples it at sparse times as snapshots, runs scFM, and checks whether the recovered trajectories match the true paths within measurement error, that would confirm or refute the claim.

Figures

Figures reproduced from arXiv: 2605.22340 by Haotian Chen, Hengshu Zhu, Jiajia Wang, Meng Xiao, Qingqing Long, Siyu Pu, Xiaohan Huang, Xiao Luo, Xuezhi Wang, Yuanchun Zhou.

Figure 1
Figure 1. Figure 1: Given sparsely sampled scRNA-seq snapshots, scFM [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: scFM models continuous-time single-cell dynamics in a VAE latent space using a two-stage training framework. (1) [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Averaged Wasserstein and ℓ2 distance between true and predicted expressions on unmeasured timepoints [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Trajectory recovery on Zebrafish under the remove [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: 2D UMAP visualization of true and predicted ex [PITH_FULL_IMAGE:figures/full_fig_p008_5.png] view at source ↗
read the original abstract

Single-cell RNA sequencing (scRNA-seq) provides high-dimensional profiles of cellular states, enabling data-driven modeling of cellular dynamics over time. In practice, time-resolved scRNA-seq is collected at only a few discrete time points as unpaired snapshot populations, leaving substantial temporal gaps. This motivates trajectory inference at unmeasured time points. Existing methods mainly follow two directions, optimal-transport (OT) alignment provides distribution-level matching between observed snapshots, while continuous-time generative models support forecasting via learned dynamics. However, two challenges remain: (i) unpaired snapshots render local transitions between adjacent time points ambiguous, leading to unstable supervision; and (ii) long-horizon prediction relies on repeated integration, where small modeling errors compound and cause distribution drift. To address these challenges, we propose single-cell Flow Matching (scFM), a latent generative framework based on coupling-conditioned flow matching. First, we compute entropically regularized OT couplings between adjacent snapshots and use them to construct soft, weighted flow-matching targets for learning time-dependent velocity fields. Second, we learn bidirectional velocity fields and leverage their consistency to refine couplings and improve temporal coherence under sparse supervision. Third, we introduce distribution-level alignment and latent dynamic regularization to anchor long rollouts and mitigate drift. Experiments on real-world time-series scRNA-seq datasets show that scFM consistently improves distributional prediction performance for both temporal interpolation and extrapolation. Moreover, scFM yields more accurate trajectory reconstruction and temporally coherent visualizations where intermediate time points are absent, indicating a more faithful recovery of underlying temporal gene expression dynamics.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 2 minor

Summary. The paper introduces single-cell Flow Matching (scFM), a latent generative model that computes entropically regularized OT couplings between adjacent unpaired scRNA-seq snapshots to form soft weighted targets for conditional flow matching of time-dependent velocity fields. Bidirectional consistency is used to refine the couplings, while distribution alignment and latent dynamic regularization are added to reduce drift during long-horizon rollouts. Experiments on real time-series scRNA-seq datasets are reported to show gains in distributional prediction for interpolation/extrapolation and in trajectory reconstruction.

Significance. If the reported improvements are robust, the work offers a concrete way to combine OT-based alignment with continuous flow matching for trajectory inference from snapshot data, addressing a common limitation in scRNA-seq where no ground-truth paths exist. The bidirectional consistency mechanism and explicit drift-mitigation terms are potentially reusable ideas for other unpaired time-series settings.

major comments (3)
  1. [§3.2] §3.2 (OT coupling construction): the claim that entropically regularized OT provides stable soft targets for velocity-field learning is load-bearing, yet the manuscript does not quantify how sensitive the downstream interpolation/extrapolation metrics are to the choice of regularization strength; an ablation varying this hyper-parameter and reporting effect on Wasserstein distance or trajectory error would be required to substantiate stability.
  2. [§4.3] §4.3 (long-rollout experiments): the bidirectional consistency loss is presented as mitigating distribution drift, but no quantitative comparison is given between rollouts with and without the consistency term on the extrapolation task; without this control it is difficult to isolate whether the reported gains come from the flow-matching objective or from the post-hoc fixes.
  3. [Table 2] Table 2 (trajectory reconstruction metrics): the reported improvements over baselines are modest in some datasets; if the difference falls within the variability of the OT coupling itself, the central claim that scFM recovers more faithful dynamics would need stronger statistical support (e.g., paired significance tests across multiple random seeds).
minor comments (2)
  1. [Eq. 5] The notation for the conditional flow-matching objective (Eq. 5) mixes time index t with the learned velocity v_θ; a short appendix clarifying the conditioning variables would improve readability.
  2. [Figure 3] Figure 3 (visualizations) lacks error bands on the interpolated distributions; adding them would make the qualitative claim of temporal coherence easier to assess.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive comments. We respond to each major comment below and indicate the revisions we will make to the manuscript.

read point-by-point responses

  1. Referee: [§3.2] §3.2 (OT coupling construction): the claim that entropically regularized OT provides stable soft targets for velocity-field learning is load-bearing, yet the manuscript does not quantify how sensitive the downstream interpolation/extrapolation metrics are to the choice of regularization strength; an ablation varying this hyper-parameter and reporting effect on Wasserstein distance or trajectory error would be required to substantiate stability.

    Authors: We agree that quantifying the sensitivity to the regularization strength is important for substantiating our claims. We will perform and report an ablation study in the revised manuscript, varying the entropic regularization parameter and showing its effects on the Wasserstein distance and trajectory error metrics for interpolation and extrapolation. revision: yes

  2. Referee: [§4.3] §4.3 (long-rollout experiments): the bidirectional consistency loss is presented as mitigating distribution drift, but no quantitative comparison is given between rollouts with and without the consistency term on the extrapolation task; without this control it is difficult to isolate whether the reported gains come from the flow-matching objective or from the post-hoc fixes.

    Authors: We thank the referee for this suggestion. To isolate the effect of the bidirectional consistency loss, we will add a control experiment in the revision comparing the extrapolation performance with and without this term, providing quantitative evidence of its contribution to reducing drift in long-horizon predictions. revision: yes

  3. Referee: [Table 2] Table 2 (trajectory reconstruction metrics): the reported improvements over baselines are modest in some datasets; if the difference falls within the variability of the OT coupling itself, the central claim that scFM recovers more faithful dynamics would need stronger statistical support (e.g., paired significance tests across multiple random seeds).

    Authors: We recognize that the improvements are modest in some cases and that additional statistical analysis would strengthen the claims. In the revised manuscript, we will include paired significance tests across multiple random seeds and report p-values in Table 2 to demonstrate that the observed differences are statistically significant beyond the variability of the OT couplings. revision: yes

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The paper introduces scFM as a constructive method that first computes entropically regularized OT couplings from unpaired snapshots to form soft targets, then augments conditional flow matching with independent bidirectional consistency refinement and distribution-level regularization terms. These steps add new supervisory signals and constraints rather than re-expressing the final velocity field or performance metric as a direct function of the initial OT couplings by construction. No self-citations, uniqueness theorems, or ansatzes from prior author work are invoked to force the central claims, and evaluation occurs on held-out real-world datasets. The derivation chain therefore remains self-contained with external content.

Axiom & Free-Parameter Ledger

2 free parameters · 1 axioms · 0 invented entities

The approach relies on several modeling choices whose values are not derived from first principles and must be selected or tuned for the data.

free parameters (2)
  • entropic regularization strength in OT
    Controls softness of couplings between snapshots; value chosen to balance matching quality and stability.
  • weights for distribution alignment and latent dynamic regularization
    Hyperparameters introduced to anchor long-horizon rollouts; not fixed by theory.
axioms (1)
  • domain assumption OT couplings between adjacent snapshots provide usable soft supervision for velocity field learning despite unpaired cells
    Invoked to address ambiguity in local transitions.

pith-pipeline@v0.9.0 · 5837 in / 1352 out tokens · 53453 ms · 2026-05-22T07:08:10.787686+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

51 extracted references · 51 canonical work pages · 4 internal anchors

  1. [1]

    Volker Bergen, Marius Lange, Stefan Peidli, F Alexander Wolf, and Fabian J Theis

    work page

  2. [2]

    Generalizing RNA velocity to transient cell states through dynamical modeling.Nature biotechnology38, 12 (2020), 1408–1414

    work page 2020

  3. [3]

    Charlotte Bunne, Geoffrey Schiebinger, Andreas Krause, Aviv Regev, and Marco Cuturi. 2024. Optimal transport for single-cell and spatial omics.Nature Reviews Methods Primers4, 1 (2024), 58

    work page 2024

  4. [4]

    Diego Calderon, Ronnie Blecher-Gonen, Xingfan Huang, Stefano Secchia, James Kentro, Riza M Daza, Beth Martin, Alessandro Dulja, Christoph Schaub, Cole Trapnell, et al. 2022. The continuum of Drosophila embryonic development at single-cell resolution.Science377, 6606 (2022), eabn5800

    work page 2022

  5. [5]

    Ricky TQ Chen, Yulia Rubanova, Jesse Bettencourt, and David K Duvenaud. 2018. Neural ordinary differential equations.Advances in neural information processing systems31 (2018)

    work page 2018

  6. [6]

    Lenaic Chizat, Gabriel Peyré, Bernhard Schmitzer, and François-Xavier Vialard

    work page

  7. [7]

    Journal of Functional Analysis274, 11 (2018), 3090–3123

    Unbalanced optimal transport: Dynamic and Kantorovich formulations. Journal of Functional Analysis274, 11 (2018), 3090–3123

    work page 2018

  8. [8]

    Marco Cuturi. 2013. Sinkhorn distances: Lightspeed computation of optimal transport.Advances in neural information processing systems26 (2013)

    work page 2013

  9. [9]

    Jun Ding, Nadav Sharon, and Ziv Bar-Joseph. 2022. Temporal modelling using single-cell transcriptomics.Nature Reviews Genetics23, 6 (2022), 355–368

    work page 2022

  10. [10]

    Florian Erhard, Antoine-Emmanuel Saliba, Alexandra Lusser, Christophe Tous- saint, Thomas Hennig, Bhupesh K Prusty, Daniel Kirschenbaum, Kathleen Abadie, Eric A Miska, Caroline C Friedel, et al. 2022. Time-resolved single-cell RNA-seq using metabolic RNA labelling.Nature Reviews Methods Primers2, 1 (2022), 77

    work page 2022

  11. [11]

    Zheng Fang, Qingqing Long, Guojie Song, and Kunqing Xie. 2021. Spatial- temporal graph ode networks for traffic flow forecasting. InProceedings of the 27th ACM SIGKDD conference on knowledge discovery & data mining. 364–373

    work page 2021

  12. [12]

    Jeffrey A Farrell, Yiqun Wang, Samantha J Riesenfeld, Karthik Shekhar, Aviv Regev, and Alexander F Schier. 2018. Single-cell reconstruction of developmental trajectories during zebrafish embryogenesis.Science360, 6392 (2018), eaar3131

    work page 2018

  13. [13]

    Jean Feydy, Thibault Séjourné, François-Xavier Vialard, Shun-ichi Amari, Alain Trouvé, and Gabriel Peyré. 2019. Interpolating between optimal transport and mmd using sinkhorn divergences. InThe 22nd international conference on artificial intelligence and statistics. PMLR, 2681–2690

    work page 2019

  14. [14]

    Rémi Flamary, Nicolas Courty, Alexandre Gramfort, Mokhtar Z Alaya, Aurélie Boisbunon, Stanislas Chambon, Laetitia Chapel, Adrien Corenflos, Kilian Fatras, Nemo Fournier, et al. 2021. Pot: Python optimal transport.Journal of Machine Learning Research22, 78 (2021), 1–8

    work page 2021

  15. [15]

    Adam Gayoso, Philipp Weiler, Mohammad Lotfollahi, Dominik Klein, Justin Hong, Aaron M Streets, Fabian J Theis, and Nir Yosef. 2022. Deep generative modeling of transcriptional dynamics for rna velocity analysis in single cells. bioRxiv.Published online January1 (2022), 2022–08

    work page 2022

  16. [16]

    Monica Golumbeanu, Sara Cristinelli, Sylvie Rato, Miguel Munoz, Matthias Cavassini, Niko Beerenwinkel, and Angela Ciuffi. 2018. Single-cell RNA-Seq reveals transcriptional heterogeneity in latent and reactivated HIV-infected cells. Cell reports23, 4 (2018), 942–950

    work page 2018

  17. [17]

    Will Grathwohl, Ricky TQ Chen, Jesse Bettencourt, Ilya Sutskever, and David Duvenaud. 2018. Ffjord: Free-form continuous dynamics for scalable reversible generative models.arXiv preprint arXiv:1810.01367(2018)

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  18. [18]

    Wenbo Guo, Zeyu Chen, and Jin Gu. 2025. Computational modeling of single-cell dynamics data.Briefings in Bioinformatics26, 3 (2025), bbaf305

    work page 2025

  19. [19]

    Laleh Haghverdi, Maren Büttner, F Alexander Wolf, Florian Buettner, and Fabian J Theis. 2016. Diffusion pseudotime robustly reconstructs lineage branching.Na- ture methods13, 10 (2016), 845–848

    work page 2016

  20. [20]

    Ashraful Haque, Jessica Engel, Sarah A Teichmann, and Tapio Lönnberg. 2017. A practical guide to single-cell RNA-sequencing for biomedical research and clinical applications.Genome medicine9, 1 (2017), 75

    work page 2017

  21. [21]

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. 2020. Denoising diffusion probabilistic models.Advances in neural information processing systems33 (2020), 6840–6851

    work page 2020

  22. [22]

    Guillaume Huguet, Daniel Sumner Magruder, Alexander Tong, Oluwadamilola Fasina, Manik Kuchroo, Guy Wolf, and Smita Krishnaswamy. 2022. Manifold interpolating optimal-transport flows for trajectory inference.Advances in neural information processing systems35 (2022), 29705–29718

    work page 2022

  23. [23]

    Diederik P Kingma and Max Welling. 2013. Auto-encoding variational bayes. arXiv preprint arXiv:1312.6114(2013)

    work page internal anchor Pith review Pith/arXiv arXiv 2013

  24. [24]

    Dominik Klein, Giovanni Palla, Marius Lange, Michal Klein, Zoe Piran, Manuel Gander, Laetitia Meng-Papaxanthos, Michael Sterr, Lama Saber, Changying Jing, et al. 2025. Mapping cells through time and space with moscot.Nature638, 8052 (2025), 1065–1075

    work page 2025

  25. [25]

    Gioele La Manno, Ruslan Soldatov, Amit Zeisel, Emelie Braun, Hannah Hochgerner, Viktor Petukhov, Katja Lidschreiber, Maria E Kastriti, Peter Lönner- berg, Alessandro Furlan, et al. 2018. RNA velocity of single cells.Nature560, 7719 (2018), 494–498

    work page 2018

  26. [26]

    Marius Lange, Volker Bergen, Michal Klein, Manu Setty, Bernhard Reuter, Mostafa Bakhti, Heiko Lickert, Meshal Ansari, Janine Schniering, Herbert B Schiller, et al

    work page

  27. [27]

    CellRank for directed single-cell fate mapping.Nature methods19, 2 (2022), 159–170

    work page 2022

  28. [28]

    Matthias Liero, Alexander Mielke, and Giuseppe Savaré. 2018. Optimal entropy- transport problems and a new Hellinger–Kantorovich distance between positive measures.Inventiones mathematicae211, 3 (2018), 969–1117

    work page 2018

  29. [29]

    Yue Ling, Peiqi Zhang, Zhenyi Zhang, and Peijie Zhou. 2025. CellStream: Dy- namical Optimal Transport Informed Embeddings for Reconstructing Cellular Trajectories from Snapshots Data.arXiv preprint arXiv:2511.13786(2025)

    work page arXiv 2025

  30. [30]

    Yaron Lipman, Ricky TQ Chen, Heli Ben-Hamu, Maximilian Nickel, and Matt Le

    work page

  31. [31]

    Flow matching for generative modeling.arXiv preprint arXiv:2210.02747 (2022)

    work page internal anchor Pith review Pith/arXiv arXiv 2022

  32. [32]

    Vivien Marx. 2024. scRNA-seq: oh, the joys.Nature Methods21, 5 (2024), 750–753

    work page 2024

  33. [33]

    Gabriel Peyré, Marco Cuturi, et al. 2019. Computational optimal transport: With applications to data science.Foundations and Trends®in Machine Learning11, 5-6 (2019), 355–607

    work page 2019

  34. [34]

    Aram-Alexandre Pooladian, Heli Ben-Hamu, Carles Domingo-Enrich, Brandon Amos, Yaron Lipman, and Ricky TQ Chen. 2023. Multisample flow matching: Straightening flows with minibatch couplings.arXiv preprint arXiv:2304.14772 (2023)

    work page arXiv 2023

  35. [35]

    Martin Rohbeck, Edward De Brouwer, Charlotte Bunne, Jan-Christian Huetter, Anne Biton, Kelvin Y Chen, Aviv Regev, and Romain Lopez. 2025. Modeling complex system dynamics with flow matching across time and conditions. In The Thirteenth International Conference on Learning Representations

    work page 2025

  36. [36]

    Wouter Saelens, Robrecht Cannoodt, Helena Todorov, and Yvan Saeys. 2019. A comparison of single-cell trajectory inference methods.Nature biotechnology37, 5 (2019), 547–554

    work page 2019

  37. [37]

    Antoine-Emmanuel Saliba, Alexander J Westermann, Stanislaw A Gorski, and Jörg Vogel. 2014. Single-cell RNA-seq: advances and future challenges.Nucleic acids research42, 14 (2014), 8845–8860

    work page 2014

  38. [38]

    Geoffrey Schiebinger, Jian Shu, Marcin Tabaka, Brian Cleary, Vidya Subramanian, Aryeh Solomon, Joshua Gould, Siyan Liu, Stacie Lin, Peter Berube, et al . 2019. Optimal-transport analysis of single-cell gene expression identifies developmen- tal trajectories in reprogramming.Cell176, 4 (2019), 928–943

    work page 2019

  39. [39]

    Yutong Sha, Yuchi Qiu, Peijie Zhou, and Qing Nie. 2024. Reconstructing growth and dynamic trajectories from single-cell transcriptomics data.Nature Machine Intelligence6, 1 (2024), 25–39

    work page 2024

  40. [40]

    K Street, D Risso, RB Fletcher, D Das, J Ngai, N Yosef, E Purdom, and SS Dudoit

    work page

  41. [41]

    Cell lineage and pseudotime inference for single-cell transcriptomics.BMC genomics19 (2018), 477

    work page 2018

  42. [42]

    Alexander Tong, Kilian Fatras, Nikolay Malkin, Guillaume Huguet, Yanlei Zhang, Jarrid Rector-Brooks, Guy Wolf, and Yoshua Bengio. 2023. Improving and gener- alizing flow-based generative models with minibatch optimal transport.arXiv preprint arXiv:2302.00482(2023). From Snapshots to Trajectories: Learning Single-Cell Gene Expression Dynamics via Condition...

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  43. [43]

    Alexander Tong, Jessie Huang, Guy Wolf, David Van Dijk, and Smita Krish- naswamy. 2020. Trajectorynet: A dynamic optimal transport network for mod- eling cellular dynamics. InInternational conference on machine learning. PMLR, 9526–9536

    work page 2020

  44. [44]

    Cole Trapnell, Davide Cacchiarelli, Jonna Grimsby, Prapti Pokharel, Shuqiang Li, Michael Morse, Niall J Lennon, Kenneth J Livak, Tarjei S Mikkelsen, and John L Rinn. 2014. The dynamics and regulators of cell fate decisions are revealed by pseudotemporal ordering of single cells.Nature biotechnology32, 4 (2014), 381–386

    work page 2014

  45. [45]

    Dongyi Wang, Yuanwei Jiang, Zhenyi Zhang, Xiang Gu, Peijie Zhou, and Jian Sun

    work page

  46. [46]

    arXiv preprint arXiv:2505.13413(2025)

    Joint Velocity-Growth Flow Matching for Single-Cell Dynamics Modeling. arXiv preprint arXiv:2505.13413(2025)

    work page arXiv 2025

  47. [47]

    Grace Hui Ting Yeo, Sachit D Saksena, and David K Gifford. 2021. Generative modeling of single-cell time series with PRESCIENT enables prediction of cell trajectories with interventions.Nature communications12, 1 (2021), 3222

    work page 2021

  48. [48]

    Kun Yin, Yiling Xu, Ye Guo, Zhong Zheng, Xinrui Lin, Meijuan Zhao, He Dong, Dianyi Liang, Zhi Zhu, Junhua Zheng, et al. 2024. Dyna-vivo-seq unveils cellular RNA dynamics during acute kidney injury via in vivo metabolic RNA labeling- based scRNA-seq.Nature Communications15, 1 (2024), 9866

    work page 2024

  49. [49]

    Jiaqi Zhang, Erica Larschan, Jeremy Bigness, and Ritambhara Singh. 2024. scN- ODE: generative model for temporal single cell transcriptomic data prediction. Bioinformatics40, Supplement_2 (2024), ii146–ii154

    work page 2024

  50. [50]

    Xiaowen Zhang, Mingjian Peng, Jianghao Zhu, Xue Zhai, Chaoguang Wei, He Jiao, Zhichao Wu, Songqian Huang, Mingli Liu, Wenhao Li, et al. 2025. Bench- marking metabolic RNA labeling techniques for high-throughput single-cell RNA sequencing.Nature Communications16, 1 (2025), 5952

    work page 2025

  51. [51]

    Zhenyi Zhang, Tiejun Li, and Peijie Zhou. 2024. Learning stochastic dynamics from snapshots through regularized unbalanced optimal transport.arXiv preprint arXiv:2410.00844(2024)

    work page arXiv 2024