Pith. sign in

REVIEW 3 major objections 5 minor 49 references

LapDDPM: A Conditional Graph Diffusion Model for scRNA-seq Generation with Spectral Adversarial Perturbations

T0 review · 3 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read LapDDPM, a conditional graph diffusion model for scRNA-seq generation, claims a new benchmark in matching real single-cell distributions.

desk verdict A sensible but incremental scRNA-seq generator whose central robustness claim is promised but never tested; the clean-graph results are decent, but the paper needs major revision. read the letter →

arxiv 2506.13344 v1 pith:4NGC2XSN submitted 2025-06-16 cs.LG cs.AIq-bio.BMq-bio.CBq-bio.GN

Lorenzo Bini , Stephane Marchand-Maillet This is my paper · ORCID
classification cs.LGcs.AIq-bio.BMq-bio.CBq-bio.GN
keywords single-cellRNA-seqgenerationconditionaldiffusionmodelsgraphneuralnetworksLaplacianpositionalencodingsspectraladversarialperturbationsmaximummeandiscrepancyWassersteindistancesyntheticdata
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

LapDDPM is a generative model for single-cell RNA sequencing data that aims to produce synthetic count matrices matching real cells, with generation controlled by cell-type labels. The paper's central claim is that combining a graph representation of cell–cell similarity with a latent score-based diffusion model, plus a spectral adversarial perturbation of graph edge weights during training, yields higher-fidelity conditional generation than existing scRNA-seq generators. On four datasets spanning roughly 2,600 to 585,000 cells, LapDDPM reports lower RBF-kernel MMD and 2-Wasserstein distances to held-out real cells than scVI, scDiffusion, and CFGen in conditional mode, and than scGAN in unconditional mode. Ablations on PBMC3K attribute part of the gain to each component: spectral adversarial perturbations, input gene masking, and Laplacian positional encodings. If these results hold, the method offers a scalable way to synthesize cell-type-specific scRNA-seq data for data augmentation, benchmarking, and in-silico experiments.

What carries the argument

The load-bearing object is the graph representation of the cell population: a k-NN graph built on PCA-reduced log-normalized expression, enriched by Laplacian positional encodings derived from the normalized graph Laplacian. The spectral encoder uses Chebyshev convolutions to embed nodes, and a conditional variance-preserving score-based diffusion model generates latent states that a Poisson decoder turns into count data. The distinctive mechanism is the spectral adversarial perturbation: a perturbation to edge weights is set proportional to products of principal-eigenvector components $v_i v_j$ of the adjacency matrix, so training sees graphs whose dominant spectral modes have been altered; this is what the paper claims supplies robustness to structural variation.

What would settle it

Compute marker-gene recall and cell-type separation on generated profiles: if synthetic cells for a known type do not express its canonical markers, or if their clusters overlap real cells of other types, the paper's claim of biologically plausible cell-type-specific generation is falsified. Similarly, a graph attack that randomly rewires edges or removes hub nodes would falsify the robustness claim if generation quality collapses.

Watch

Extended reading notes

Core claim

LapDDPM is a conditional generative model that maps gene expression counts into a latent space through a Chebyshev graph convolutional encoder operating on a k-NN cell graph augmented with Laplacian positional encodings, then learns a score-based diffusion process in that latent space conditioned on cell-type labels, and decodes samples back into gene expression rates via a Poisson likelihood. The paper reports that this system achieves lower RBF-kernel MMD and 2-Wasserstein distances to held-out real cells than scVI, scDiffusion, and CFGen for conditional generation, and lower unconditional distances than scGAN, across PBMC3K, Dentate Gyrus, Tabula Muris, and HLCA. Ablations indicate that removing the spectral adversarial perturbation, the input gene masking, or the Laplacian positional encodings each degrades these metrics, which the paper reads as evidence that the components contribute to fidelity and robustness.

Load-bearing premise

The central claim stands on the assumption that the nearest-neighbour graph built from the leading principal components of log-normalized expression faithfully represents cell–cell relationships, and that matching real data on two distribution-distance metrics in a 30-dimensional projection is enough to certify biological plausibility and robustness.

Editorial extensions

If this is right

  • Conditional generation can supply synthetic cells for rare cell types whose real counts are too low for training or testing.
  • The latent-space diffusion setup with a Poisson decoder gives a recipe for generating count matrices rather than normalized expression only, which is what many downstream single-cell tools expect.
  • The ablations imply each architectural choice (LPEs, input masking, spectral perturbation) is doing measurable work, so future generators in this line should keep or replace them deliberately.
  • Because preprocessing, training, and generation scale approximately linearly in the number of cells, the method is positioned for atlas-scale datasets such as HLCA.
  • Robustness to perturbed edge weights means the generator could tolerate noisy cell graphs from fresh experiments or different preprocessing pipelines.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Because the paper evaluates only distribution distances in PCA space, the more consequential claim of biological plausibility remains untested; marker-gene, trajectory, and differential-expression checks would be the natural next experiments.
  • The spectral adversarial perturbation has the flavor of a spectral data augmentation and could plausibly transfer to other graph-conditioned generative tasks beyond single cells, such as spatial transcriptomics or graph-based drug response modeling, whenever the graph is noisy.
  • If the k-NN graph is itself a noisy estimate of cell state, the robustness conferred by training on perturbed graphs may also make the encoder less sensitive to the choice of k and PCA dimension, which would simplify practical deployment.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. LapDDPM is a conditional graph diffusion model for generating synthetic scRNA-seq data. The method builds a k-NN graph on PCA-reduced, log-normalized gene expression, augments cell features with Laplacian positional encodings, and trains a latent score-based diffusion model conditioned on cell-type or tissue labels. A spectral adversarial perturbation module modifies graph edge weights based on the principal eigenvector of the adjacency matrix during encoder training, ostensibly to improve robustness. The paper evaluates conditional and unconditional generation on PBMC3K, Dentate Gyrus, Tabula Muris, and HLCA using RBF-kernel MMD and 2-Wasserstein distances in a 30-dimensional PCA projection, and reports ablations of the main components.

Significance. If the proposed robustness mechanism worked as intended, the work would be a useful contribution to conditional scRNA-seq generation, combining graph structure with diffusion models. The paper offers a broad set of baselines and ablations, and the clean-graph results are competitive. However, the central novelty—spectral adversarial perturbation for robustness—is never directly evaluated, and the biological plausibility claim rests on distribution distances alone. As it stands, the evidence does not support the abstract's claims of a 'robust tool' or a 'new benchmark'; the paper would become significantly stronger if the authors supplied the promised adversarial experiments and biological validation.

major comments (3)
  1. [Section 2.4, Section 4, Table 3] Section 2.4 promises an evaluation in Section 4 of robustness against structural attacks (random, DICE, GF-Attack, Mettack), but Section 4 contains no such experiments; Table 1 and Table 3 report only clean-graph MMD and 2-Wasserstein distances. Since the spectral perturbation is the paper's distinguishing contribution, the robustness claim is unsupported. Moreover, because the graph is used only during encoder training and not at generation time, the mechanism by which this training-time perturbation would confer test-time robustness is not explained; the authors should add the promised evaluation or revise the claims.
  2. [Section 4.1] The quantitative evaluation is limited to RBF-kernel MMD and 2-Wasserstein distance on 30-dimensional PCA projections. No biological validation (e.g., marker-gene expression, cluster structure, differential expression, or downstream task performance) is provided, so the claims of generating 'biologically-plausible, cell-type-specific samples' are not substantiated.
  3. [Table 1] Several entries show overlapping error bars with the strongest baseline, such as Tabula Muris conditional MMD (LapDDPM 0.19±0.02 vs CFGen 0.19±0.02) and Dentate Gyrus conditional MMD (LapDDPM 1.04±0.08 vs CFGen 1.12±0.04). No significance tests are reported over the 10 runs, so the statement that LapDDPM 'consistently' outperforms baselines is not fully supported.
minor comments (5)
  1. [Section 3.4] The text calls wdiff, wKL, and wrec 'learnable weights,' but Algorithm 1 lists them as fixed input hyperparameters; please clarify whether they are optimized during training or hand-tuned.
  2. [Section 2.4] Mettack is cited as (Gosch et al., 2024b), but Mettack originates from Zügner et al. (2018), which is already cited in the same paragraph; please correct this reference.
  3. [Table 1] The abbreviations 'C-CFG EN' and 'U-CFG EN' are unclear; please rename them to 'CFGen (conditional)' and 'CFGen (unconditional)' for readability.
  4. [Section 5.2, Table 3] The text states that the ablation 'assesses ... robustness to structural variations,' but Table 3 only reports clean-graph MMD/WD; the wording should match the actual experiment.
  5. [General] The paper does not state whether code is publicly available; providing a repository would improve reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: the reported MMD and 2-Wasserstein improvements are evaluated on held-out data and do not reduce to the training inputs; the unfulfilled robustness-evaluation promise is an evidentiary gap, not circularity.

full rationale

The derivation chain of LapDDPM is self-contained against external benchmarks. The model is trained with a combined diffusion, KL, and Poisson-NLL reconstruction loss (Section 3.4), and the reported MMD and 2-Wasserstein metrics are computed on held-out test data by embedding generated samples into a PCA space fitted to real test data (Section 4.1); no equation defines these metrics as a training objective or as a function of a fitted parameter. The k-NN graph and Laplacian positional encodings are constructed from the expression data (Section 3.1), and the spectral adversarial perturbation is a fixed transformation of the graph's principal eigenvector (Section 3.3), so it is not an inversion of the evaluation metric. The paper contains no load-bearing self-citation: all cited prior methods are external baselines. The central robustness claim is weakened by the paper's own internal promise in Section 2.4 that Section 4 would assess adversarially poisoned input graphs, which is not delivered by the clean-graph experiments in Table 1 and Table 3; that is a missing-evidence and correctness concern, not a circular-reasoning defect. Under the required standard that circularity be exhibited as a specific reduction, no circular step is present.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The central claims rest on standard machine-learning assumptions (latent space Gaussianity, Poisson count likelihood) and on the unvalidated premise that a k-NN graph in PCA space plus the reported distribution distances capture biological fidelity. No invented entities are introduced.

free parameters (4)
  • Loss weights wdiff, wKL, wrec = not specified (described as learnable but no update rule given)
    Section 3.4: LTotal = wdiff*Ldiff + wKL*LKL + wrec*Lrec; the text calls them learnable weights, but Algorithm 1 treats them as inputs, so their values and update procedure are unclear.
  • Input masking fraction m = 10%-30% (from ablation title)
    Section 3.4 and Table 3: the masking fraction is a data-augmentation hyperparameter; the ablation only reports a range, not the exact value used for the main results.
  • Spectral perturbation parameters αmin, αmax, ε, ip = not specified
    Section 3.3: the perturbation sampling range, scaling factor, and power-iteration count are central to the novelty claim but no values are given.
  • Graph construction hyperparameters (k-NN k, PCA dim, LPE dim) = not specified
    Section 3.1: the k-NN graph, PCA dimension, and LPE eigenvector count are not reported, so the main results cannot be exactly reproduced.
assumptions (4)
  • domain assumption A k-NN graph built on PCA-reduced, log-normalized gene expression captures biologically meaningful cell-cell relationships.
    Section 3.1: the entire graph-based representation, LPEs, and spectral perturbations depend on this graph being a faithful cellular similarity structure; the paper does not validate this against biology.
  • domain assumption The latent space produced by the graph encoder is approximately Gaussian and is a valid space for a VP-SDE score-based diffusion model with a standard normal prior.
    Section 3.2: the model assumes z0 ~ N(µ, exp(log σ²)) and samples zT ~ N(0, I); the KL term encourages this, but there is no check that the learned latent distribution is smooth enough for diffusion sampling to yield meaningful z0.
  • domain assumption RBF-kernel MMD and 2-Wasserstein distance on 30-dimensional PCA projections of real vs generated data are valid proxies for generation fidelity and biological plausibility.
    Section 4.1: these are the only quantitative fidelity metrics; no biological validation, such as marker gene expression, differential expression, or downstream clustering, is reported, so the biologically plausible claim relies entirely on this assumption.
  • domain assumption Poisson likelihood with decoder log-rates is an appropriate generative model for scRNA-seq counts after gene filtering.
    Section 3.2: the decoder outputs log(rates) and counts are Poisson-sampled; negative-binomial or zero-inflated models are standard for scRNA-seq dropouts, and the paper does not justify Poisson over these alternatives.

how reviews work

0 comments
Cite this review

Pith. "Pith review of LapDDPM: A Conditional Graph Diffusion Model for scRNA-seq Generation with Spectral Adversarial Perturbations." pith.science (2026). https://pith.science/paper/4NGC2XSN

@misc{pith2026250613344,
  author       = {Pith},
  title        = {Pith review of: LapDDPM: A Conditional Graph Diffusion Model for scRNA-seq Generation with Spectral Adversarial Perturbations},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4NGC2XSN}},
  note         = {Machine review of arXiv:2506.13344}
}
read the original abstract

Generating high-fidelity and biologically plausible synthetic single-cell RNA sequencing (scRNA-seq) data, especially with conditional control, is challenging due to its high dimensionality, sparsity, and complex biological variations. Existing generative models often struggle to capture these unique characteristics and ensure robustness to structural noise in cellular networks. We introduce LapDDPM, a novel conditional Graph Diffusion Probabilistic Model for robust and high-fidelity scRNA-seq generation. LapDDPM uniquely integrates graph-based representations with a score-based diffusion model, enhanced by a novel spectral adversarial perturbation mechanism on graph edge weights. Our contributions are threefold: we leverage Laplacian Positional Encodings (LPEs) to enrich the latent space with crucial cellular relationship information; we develop a conditional score-based diffusion model for effective learning and generation from complex scRNA-seq distributions; and we employ a unique spectral adversarial training scheme on graph edge weights, boosting robustness against structural variations. Extensive experiments on diverse scRNA-seq datasets demonstrate LapDDPM's superior performance, achieving high fidelity and generating biologically-plausible, cell-type-specific samples. LapDDPM sets a new benchmark for conditional scRNA-seq data generation, offering a robust tool for various downstream biological applications.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

49 extracted references · 39 canonical work pages

  1. [1]

    PBMC3K dataset, 2017

    10x Genomics . PBMC3K dataset, 2017. A widely used single-cell RNA sequencing dataset from 3k Peripheral Blood Mononuclear Cells

    work page 2017

  2. [2]

    Protein diffusion models for generation and design

    Anand, N., Achim, K., Eguchi, R., Ho, J., Kim, Y.-S., Li, Y., and Rives, A. Protein diffusion models for generation and design. Advances in Neural Information Processing Systems, 35: 0 2721--2733, 2022

    work page 2022

  3. [3]

    Analyzing the expressive power of graph neural networks in a spectral perspective

    Balcilar, M., Renton, G., H \'e roux, P., Ga \"u z \`e re, B., Adam, S., and Honeine, P. Analyzing the expressive power of graph neural networks in a spectral perspective. In International Conference on Learning Representations, 2021

    work page 2021

  4. [4]

    Graph contrastive learning with stable and scalable spectral encoding

    Bo, D., Fang, Y., Liu, Y., and Shi, C. Graph contrastive learning with stable and scalable spectral encoding. Advances in Neural Information Processing Systems, 36, 2024

    work page 2024

  5. [5]

    M., Gretton, A., Sch \"o lkopf, B., Smola, A

    Borgwardt, K. M., Gretton, A., Sch \"o lkopf, B., Smola, A. J., Kriegel, H.-P., and Vishwanathan, S. V. N. Integrating knowledge from heterogeneous data sources: an application to protein-protein interaction prediction. In Proceedings of the 2006 ACM symposium on Applied computing, pp.\ 535--539, 2006

    work page 2006

  6. [6]

    A restricted black-box adversarial framework towards attacking graph embedding models

    Chang, H., Rong, Y., Xu, T., Huang, W., Zhang, H., Cui, P., Zhu, W., and Huang, J. A restricted black-box adversarial framework towards attacking graph embedding models. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 34, pp.\ 3389--3396, 2020

    work page 2020

  7. [7]

    Flow-sc : A normalizing flow-based deep generative model for single-cell rna sequencing data

    Chen, T., Chen, J., Guo, Y., and Wang, H. Flow-sc : A normalizing flow-based deep generative model for single-cell rna sequencing data. Bioinformatics, 37 0 (23): 0 4353--4360, 2021

    work page 2021

  8. [8]

    Chung, F. R. Spectral graph theory, volume 92. American Mathematical Soc., 1997

    1997

Show all 49 references
  1. [9]

    Consortium, T. M. et al. Single-cell transcriptomics of 20 mouse organs creates a tabula muris. Nature, 562 0 (7727): 0 367--372, 2018

    2018

  2. [10]

    Adversarial attack on graph structured data

    Dai, H., Li, H., Tian, T., Huang, X., Wang, L., Zhu, J., and Song, L. Adversarial attack on graph structured data. In International conference on machine learning, pp.\ 1115--1124. PMLR, 2018

    2018

  3. [11]

    Convolutional neural networks on graphs with fast localized spectral filtering

    Defferrard, M., Bresson, X., and Vandergheynst, P. Convolutional neural networks on graphs with fast localized spectral filtering. Advances in neural information processing systems, 29, 2016

    2016

  4. [12]

    and Nichol, A

    Dhariwal, P. and Nichol, A. Diffusion models beat gans on image synthesis. Advances in Neural Information Processing Systems, 34: 0 8789--8802, 2021

    2021

  5. [13]

    NICE : Non-linear independent components estimation

    Dinh, L., Krueger, D., and Bengio, Y. NICE : Non-linear independent components estimation. arXiv preprint arXiv:1410.8516, 2014

    2014 arXiv

  6. [14]

    Density estimation using real nvp

    Dinh, L., Sohl-Dickstein, J., and Bengio, S. Density estimation using real nvp. International Conference on Learning Representations, 2017

    2017

  7. [15]

    M., Mircea, M., Mueller, N

    Eraslan, G., Simon, L. M., Mircea, M., Mueller, N. S., and Theis, F. J. Single-cell rna-seq imputation and denoising using a deep count autoencoder. Nature communications, 10 0 (1): 0 3907, 2019

    2019

  8. [16]

    Graph random neural networks for semi-supervised learning on graphs

    Feng, W., Zhang, J., Dong, Y., Han, Y., Luan, H., Xu, Q., Yang, Q., Kharlamov, E., and Tang, J. Graph random neural networks for semi-supervised learning on graphs. Advances in neural information processing systems, 33: 0 22092--22103, 2020

    2020

  9. [17]

    I., Stegle, O., Theis, F

    Gayoso, A., Lopez, R., Boyeau, P., Ch \"e neby, J., Li, M., Yosef, N., Jordan, M. I., Stegle, O., Theis, F. J., and Risso, D. scVI : Probabilistic modeling of single-cell omics data in python. Nature methods, 18 0 (2): 0 165--172, 2021

    2021

  10. [18]

    Spectral augmentations for graph contrastive learning

    Ghose, A., Zhang, Y., Hao, J., and Coates, M. Spectral augmentations for graph contrastive learning. In Ruiz, F., Dy, J., and van de Meent, J.-W. (eds.), Proceedings of The 26th International Conference on Artificial Intelligence and Statistics, volume 206 of Proceedings of Ma...

    2023

  11. [19]

    Generating synthetic single-cell rna-seq data using generative adversarial networks

    Ghosh, P., Datta, A., Mukhopadhyay, S., Roy, D., and Das, A. Generating synthetic single-cell rna-seq data using generative adversarial networks. Bioinformatics, 36 0 (7): 0 2119--2127, 2020

    2020

  12. [20]

    J., Shlens, J., and Szegedy, C

    Goodfellow, I. J., Shlens, J., and Szegedy, C. Explaining and harnessing adversarial examples. arXiv preprint arXiv:1412.6572, 2014

    2014 arXiv

  13. [21]

    u gner, D., and G \

    Gosch, L., Geisler, S., Sturm, D., Charpentier, B., Z \"u gner, D., and G \"u nnemann, S. Adversarial training for graph neural networks: Pitfalls, solutions, and new directions. Advances in Neural Information Processing Systems, 36, 2024 a

    2024

  14. [22]

    u gner, D., and G \

    Gosch, L., Geisler, S., Sturm, D., Charpentier, B., Z \"u gner, D., and G \"u nnemann, S. Adversarial training for graph neural networks: Pitfalls, solutions, and new directions. Advances in Neural Information Processing Systems, 36, 2024 b

    2024

  15. [23]

    Denoising diffusion probabilistic models

    Ho, J., Jain, A., and Abbeel, P. Denoising diffusion probabilistic models. Advances in Neural Information Processing Systems, 33: 0 6840--6851, 2020

    2020

  16. [24]

    G., Vignac, C., and Welling, M

    Hoogeboom, E., Satorras, V. G., Vignac, C., and Welling, M. Equivariant diffusion for molecule generation in 3d. International Conference on Machine Learning, 2022

    2022

  17. [25]

    Y., Andrews, T

    Kiselev, V. Y., Andrews, T. S., and Hemberg, M. Challenges in computational analysis of single-cell rna sequencing data. Nature Reviews Genetics, 20 0 (10): 0 573--582, 2019

    2019

  18. [26]

    Flag: Adversarial data augmentation for graph neural networks

    Kong, K., Li, G., Ding, M., Wu, Z., Zhu, C., Ghanem, B., Taylor, G., and Goldstein, T. Flag: Adversarial data augmentation for graph neural networks. arXiv preprint arXiv:2010.09891, 2020

    2010 arXiv

  19. [27]

    Diffwave: A diffusion probabilistic model for text-to-speech

    Kong, Z., Ping, W., Huang, K., Zhao, K., Li, Z., and Catanzaro, B. Diffwave: A diffusion probabilistic model for text-to-speech. International Conference on Learning Representations, 2021

    2021

  20. [28]

    Rna velocity of single cells

    La Manno, G., Soldatov, R., Zeisel, A., Braun, E., Hochgerner, H., Mayer, F., Meletis, K., Zeisel, A., Enge, M., Betsholtz, C., et al. Rna velocity of single cells. Nature, 560 0 (7719): 0 494--498, 2018

    2018

  21. [29]

    Robust graph convolutional networks against adversarial attacks

    Li, G., Han, J., Wu, X., Tang, Y., and Liu, Z. Robust graph convolutional networks against adversarial attacks. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pp.\ 1273--1282, 2019

    2019

  22. [30]

    Adversarial neural networks for noise removal in single-cell rna-seq data

    Li, J., Zhang, P., Wu, L., Ma, S., and Chen, L. Adversarial neural networks for noise removal in single-cell rna-seq data. Bioinformatics, 37 0 (10): 0 1398--1405, 2021

    2021

  23. [31]

    DeepMoji : Vae-based generative model for text with emotional content

    Lopez, F., Al-Rubaie, M., and Al-Rubaie, R. DeepMoji : Vae-based generative model for text with emotional content. arXiv preprint arXiv:1807.03961, 2018

    2018 arXiv

  24. [32]

    scDiffusion : A diffusion model for single-cell rna sequencing data generation

    Luo, Y., Pan, Y., Li, Z., Li, Z., Wang, Z., and Zhao, Z. scDiffusion : A diffusion model for single-cell rna sequencing data generation. arXiv preprint arXiv:2401.12196, 2024

    2024 arXiv

  25. [33]

    o rnemann, J., Gribov, E., He, B., K \

    Marouf, M., Bounab, Y., D \"o rnemann, J., Gribov, E., He, B., K \"o nig, D., K \"o nig, J., K \"o nig, P., K \"o nig, T., K \"o nig, V., et al. scGAN : Generating single-cell rna-seq data with generative adversarial networks. Bioinformatics, 36 0 (15): 0 4302--4309, 2020

    2020

  26. [34]

    Multi-modal and multi-attribute generation of single cells with cfgen, 2025

    Palma, A., Richter, T., Zhang, H., Lubetzki, M., Tong, A., Dittadi, A., and Theis, F. Multi-modal and multi-attribute generation of single cells with cfgen, 2025. URL https://arxiv.org/abs/2407.11734

    2025 arXiv

  27. [35]

    Graph-based inference of cell differentiation trajectories from single-cell rna sequencing data

    Qi, J., Ma, Y., Zhang, R., Han, Z., and Li, J. Graph-based inference of cell differentiation trajectories from single-cell rna sequencing data. Genome biology, 22 0 (1): 0 1--22, 2021

    2021

  28. [36]

    High-resolution image synthesis with latent diffusion models

    Rombach, R., Blattmann, A., Lorenz, D., Esser, P., and Ommer, B. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 10684--10695, 2022

    2022

  29. [37]

    J., Gorski, S., and Vogel, J

    Saliba, A.-E., Westermann, A. J., Gorski, S., and Vogel, J. Single-cell rna-seq for pathogen-infected cells. Methods, 67 0 (3): 0 313--321, 2014

    2014

  30. [38]

    C., Gillett, T

    Sikkema, L., Ram \' rez-Su \'a stegui, C., Strobl, D. C., Gillett, T. E., Zappia, L., Madissoon, E., Markov, N. S., Zaragosi, L.-E., Ji, Y., Ansari, M., et al. An integrated cell atlas of the lung in health and disease. Nature medicine, 29 0 (6): 0 1563--1577, 2023

    2023

  31. [39]

    Deep unsupervised learning using nonequilibrium thermodynamics

    Sohl-Dickstein, J., Weiss, E., Maheswaranathan, N., and Swersky, B. Deep unsupervised learning using nonequilibrium thermodynamics. International Conference on Machine Learning, 2015

    2015

  32. [40]

    and Ermon, S

    Song, Y. and Ermon, S. Score-based generative modeling through stochastic differential equations. International Conference on Learning Representations, 2020

    2020

  33. [41]

    A., Vassar, G., Luo, Y., Gao, S., Eaton, M

    Tang, F., Barbacioru, C., Nordman, E., Li, C., Xu, N., Surani, M. A., Vassar, G., Luo, Y., Gao, S., Eaton, M. A., et al. RNA-seq analysis of single-cell transcriptome and splicing in early mouse embryos. Nature methods, 6 0 (5): 0 377--382, 2009

    2009

  34. [42]

    Grand++: Graph neural diffusion with a source term

    Thorpe, M., Nguyen, T., Xia, H., Strohmer, T., Bertozzi, A., Osher, S., and Wang, B. Grand++: Graph neural diffusion with a source term. ICLR, 2022

    2022

  35. [43]

    Graph neural networks for cell type annotation from single-cell rna sequencing data

    Wang, P., Yu, S., Wen, J., Chen, M., Sun, H., Liu, S., and Sun, J. Graph neural networks for cell type annotation from single-cell rna sequencing data. Bioinformatics, 37 0 (11): 0 1558--1566, 2021

    2021

  36. [44]

    P., Wooldridge, M

    Waniek, M., Michalak, T. P., Wooldridge, M. J., and Rahwan, T. Hiding individuals and communities in a social network. Nature Human Behaviour, 2 0 (2): 0 139--147, 2018

    2018

  37. [45]

    Graph neural networks for single-cell rna sequencing data analysis

    Zhao, Z., Li, Z., Li, Z., Wang, Z., Luo, Y., and Pan, Y. Graph neural networks for single-cell rna sequencing data analysis. Briefings in Bioinformatics, 22 0 (6): 0 bbab306, 2021

    2021

  38. [46]

    Graph convolutional neural network for spatial transcriptomics analysis

    Zhou, Y., Zhou, M., Li, R., and Liu, Y. Graph convolutional neural network for spatial transcriptomics analysis. Nature communications, 13 0 (1): 0 5550, 2022

    2022

  39. [47]

    Propagation is all you need: A new framework for representation learning and classifier training on graphs

    Zhuo, J., Cui, C., Fu, K., Niu, B., He, D., Guo, Y., Wang, Z., Wang, C., Cao, X., and Yang, L. Propagation is all you need: A new framework for representation learning and classifier training on graphs. In Proceedings of the 31st ACM International Conference on Multimedia, pp....

    2023

  40. [48]

    u gner, D., Akbarnejad, A., and G \

    Z \"u gner, D., Akbarnejad, A., and G \"u nnemann, S. Adversarial attacks on neural networks for graph data. In Proceedings of the 24th ACM SIGKDD international conference on knowledge discovery & data mining, pp.\ 2847--2856, 2018

    2018

  41. [49]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.