Pith. sign in

REVIEW 1 major objections 5 minor 135 references

Emerging AI Approaches for Cancer Spatial Omics

T0 review · 1 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read Tumor maps need AI that learns physics, not just patterns.

desk verdict A useful review that sorts spatial omics AI into three paradigms, with a speculative mechanistic agenda that slightly oversells readiness. read the letter →

arxiv 2506.23857 v1 pith:IFTVONSJ submitted 2025-06-30 q-bio.QM q-bio.TO

Javad Noorbakhsh , Ali Foroughi pour , Jeffrey Chuang This is my paper
classification q-bio.QMq-bio.TO
keywords artificialintelligencespatialtranscriptomicsproteomicsdeeplearningfoundationmodelstissuebiophysicsphysics-informedneuralnetworksinterpretableAI
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

The paper argues that cancer spatial omics has reached a point where large AI models can find patterns in tissue but cannot say why those patterns matter, and that the field's next step depends on making the models interpretable. It organizes current approaches into three paradigms: data-driven spatial AI (foundation models), constraint-based spatial AI (information theory and diffusion models), and mechanistic spatial modeling (tissue biophysics and evolution). The paper's core recommendation is that these paradigms should be integrated, with physics-informed neural networks (PINNs) and perturbable model systems such as mice playing a central role in turning spatial observations into testable biological mechanisms. If the field follows this direction, interpretability becomes a design criterion rather than an afterthought, and spatial omics can move from correlation to causal understanding of the tumor microenvironment.

What carries the argument

The organizing device is a three-part taxonomy of spatial AI paradigms, with a gradient of interpretability running from data-driven through constraint-based to mechanistic. The mechanism that carries the recommendation is the physics-informed neural network applied to tissue fields: cells are treated as relays of diffusing molecular or mechanical fields, and reaction-diffusion equations or mechanical constitutive equations are embedded in the network so that biophysical parameters (diffusion lengths, stiffness, reaction rates) become learnable outputs. Supporting machinery includes information-theoretic constraints such as the information bottleneck, diffusion models that respect spatial hierarchy, and functional tissue units as reusable structural modules.

What would settle it

Generate synthetic spatial transcriptomic data from a known reaction-diffusion system with a specified diffusion coefficient and reaction term, train a PINN on a realistic number of time points, and check whether the inferred parameters match the ground truth; if they do not, the core promise of the mechanistic paradigm is not met. Alternatively, measure a tissue property such as stiffness or ligand diffusion length experimentally and test whether PINN-inferred values from SRO data agree.

Watch

Extended reading notes

Core claim

The central claim is that the overarching challenge in cancer spatial omics is the development of interpretable spatial AI models, and that no single paradigm will suffice. Data-driven foundation models are flexible but only post-hoc interpretable; constraint-based models improve interpretability by building in biological limits such as information transfer and spatial hierarchy; mechanistic models are the most interpretable because they are constructed from experimentally perturbable entities such as cells and genes. The paper argues that physics-informed neural networks are a particularly promising mechanistic tool because they combine explicit mathematical modeling with data-driven inference, but their use on tissues is currently limited by the scarcity of time-course data. Consequently, the authors recommend purposefully generating dynamic spatial data in mouse models and building integrated human-mouse repositories to validate and transfer mechanistic knowledge.

Load-bearing premise

The recommendation assumes that sufficient dynamic or perturbable spatial omics data can be generated, especially from mouse models, and that physics-informed neural networks can infer interpretable, correct biophysical parameters from such data; the paper itself notes the current scarcity of time-course tissue data.

Editorial extensions

If this is right

  • Spatial omics foundation models should be benchmarked on multiscale, microenvironment-level tasks rather than whole-tumor classification alone.
  • Curated repositories integrating human and mouse spatial cancer data would make mechanistic validation and cross-species transfer routine.
  • PINN-style models, if supplied with mouse time-course data, could infer interpretable biophysical parameters from spatial omics data.
  • Combined multi-species tokenizers with explicit positional encodings could align human and mouse spatial data for cross-species discovery.

Reading between the lines

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

  • If the interpretability gradient is correct, a practical design rule follows: use the most mechanistic model the available data can support, reserving pure black-box models for low-data or poorly understood regimes.
  • A direct testable extension is to benchmark PINN parameter recovery on synthetic spatial omics with known ground-truth diffusion coefficients, which would quantify when the mechanistic paradigm is reliable.
  • The argument implies that investment in longitudinal and perturbable spatial atlases in model organisms may be as important as architectural advances in AI.
  • The paper's logic also suggests that steady-state approximations used by existing tools are a stopgap until true time-course data are available, so these tools could be systematically compared against full PINN solutions on matched data.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

1 major / 5 minor

Summary. This review surveys artificial intelligence approaches for cancer spatial omics, organizing the field into three paradigms: data-driven spatial AI (histopathology and spatial-omics foundation models, post-hoc interpretation), constraint-based spatial AI (information-theoretic methods, diffusion models), and mechanistic spatial modeling (tissue biophysics, evolution). The paper argues that interpretability is a central challenge, recommends integrating AI with hypothesis-driven strategies and perturbable mouse model systems, and proposes physics-informed neural networks as a promising route to mechanistic inference from spatially resolved omics data. It concludes with data integration challenges and the need for standardized benchmarking.

Significance. The paper is a timely, broad, and generally accurate synthesis with an extensive citation base through mid-2025. Its main contribution is a conceptual framework—three paradigms—that can help structure discussion and goal-setting in the field, and it explicitly identifies validation, data scarcity, and cross-species transfer as key bottlenecks. The review is strongest as a roadmap: it credits the maturity of data-driven foundation models while appropriately flagging their limited interpretability, and it advocates for model systems to generate perturbable spatial data. The mechanistic modeling proposal is clearly speculative, but the authors are transparent about the lack of time-course data. No original results are presented; the value lies in synthesis and agenda-setting.

major comments (1)
  1. [Inference of tissue biophysics (p. 10) and Conclusions (p. 14)] The review's central recommendation—that mechanistic spatial modeling with PINNs, supported by perturbable mouse SRO data, is the most promising path to interpretability—rests on the unverified premise that identifiable biophysical parameters can be recovered from SRO measurements. The paper acknowledges the scarcity of time-course data and notes that the two tissue-scale examples (HoloNet, SpaCCC) strip dynamics and use steady-state functional forms, but it does not address the identifiability problem: coupled diffusion and reaction parameters are generically non-identifiable from static spatial snapshots without perturbation or flux information, and longitudinal sampling mitigates but does not eliminate this issue. Because this feasibility premise is load-bearing for the paper's proposed research agenda, the authors should either temper the conclusion to present the PINN approach as an untested hypothesis or add a concrete discussion of identifiability challenges and possible validation benchmarks (e.g., synthetic-data recovery experiments).
minor comments (5)
  1. [Spatial omic foundation models (p. 5)] Reference [21] is cited to support the statement that tokenization specifies 'gene identity, expression value, ranking, and metadata,' but the cited paper is about tokenization for multiplex immunofluorescence and histology image synthesis, not single-cell transcriptomics; the citation should be replaced or the sentence revised.
  2. [Inference of tissue biophysics (p. 10)] The paragraph on Kolmogorov-Arnold Networks (KANs) is presented as part of mechanistic spatial modeling, but the cited remote-sensing application does not use physics-informed constraints; clarify how KANs relate to the mechanistic paradigm.
  3. [Title page] The three affiliations are all numbered '1'; the numbering should be corrected to 1, 2, 3 to distinguish the Jackson Laboratory, St. Jude, and UCONN affiliations.
  4. [Conclusions (p. 14)] The sentence 'PINNs are a particularly promising approach for mechanistic modeling' omits the caveats about time-course data scarcity stated earlier; consider adding a qualifier such as 'once sufficient longitudinal SRO data become available.'
  5. [Constraint-based spatial AI (p. 7)] The boundary between 'data-driven' and 'constraint-based' paradigms is somewhat fuzzy, as diffusion models are also trained purely from data; a sentence clarifying the distinguishing criterion would help.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper is a review/taxonomy whose recommendations rest on stated evidence gaps, not on self-referential derivation.

full rationale

This is a review and taxonomy paper whose central claim is organizational (three paradigms: data-driven spatial AI, constraint-based spatial AI, and mechanistic spatial modeling) and prescriptive (invest in mechanistic modeling and perturbable mouse SRO data). There is no derivation chain in which a predicted output is constructed from its own inputs. The self-citations ([41], [129], and possibly [104]) are used as illustrative examples of published methods or as supporting evidence for empirical claims, not as authorities that force the taxonomy or the recommendation. The paper explicitly identifies its own weakest premise: 'application of PINNs to tissues has been limited due to scarcity of time-course data.' That is an honest limitation statement and an evidence gap, not circularity: the mechanistic-modeling agenda is offered as a research direction conditional on data that do not yet exist, rather than as a result derived from those data. No equation is defined in terms of a target quantity, no fitted parameter is renamed a prediction, and no uniqueness theorem is imported from the authors' prior work to forbid alternatives. The absence of an identifiability benchmark for PINN-style inference from static SRO data is a correctness or feasibility concern, not a circularity concern. Accordingly, the appropriate finding is no significant circularity.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

No free parameters or invented entities. The review depends on the domain assumptions above rather than on fitted constants or new theoretical constructs.

assumptions (5)
  • domain assumption Spatial omics data are growing and will continue to grow; the field is early in establishing goals.
    Background section states 'The field of spatial omics is in the early stages of establishing a hierarchy of goals'; the paper's whole argument assumes data abundance will make AI approaches valuable.
  • domain assumption Interpretability is the central bottleneck for spatial AI in cancer.
    Abstract and Conclusions assert an 'overarching challenge is the development of interpretable spatial AI models'; this is a value judgment about what matters most, not a proven fact.
  • domain assumption Mechanistic models are more interpretable and preferable because they are based on experimentally perturbable entities.
    Conclusions state: 'Mechanistic models are the most interpretable, as they are constructed based on experimentally perturbable entities such as cells or genes.'
  • domain assumption Sufficient time-course/perturbable SRO data can be generated, mainly from mouse models, to train mechanistic models.
    Inference of tissue biophysics section states 'So far, application of PINNs to tissues has been limited due to scarcity of time-course data. However, as longitudinal SRO data improve...' and Mechanism-driven data generation states 'Dynamic SRO data can be generated more easily in mice than from clinical samples.'
  • standard math Attention weights do not uniquely translate to output feature importances.
    Cited from references [47]-[49]; used to justify the need for post-hoc interpretation methods.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Emerging AI Approaches for Cancer Spatial Omics." pith.science (2026). https://pith.science/paper/IFTVONSJ

@misc{pith2026250623857,
  author       = {Pith},
  title        = {Pith review of: Emerging AI Approaches for Cancer Spatial Omics},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/IFTVONSJ}},
  note         = {Machine review of arXiv:2506.23857}
}
read the original abstract

Technological breakthroughs in spatial omics and artificial intelligence (AI) have the potential to transform the understanding of cancer cells and the tumor microenvironment. Here we review the role of AI in spatial omics, discussing the current state-of-the-art and further needs to decipher cancer biology from large-scale spatial tissue data. An overarching challenge is the development of interpretable spatial AI models, an activity which demands not only improved data integration, but also new conceptual frameworks. We discuss emerging paradigms, in particular data-driven spatial AI, constraint-based spatial AI, and mechanistic spatial modeling, as well as the importance of integrating AI with hypothesis-driven strategies and model systems to realize the value of cancer spatial information.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

135 extracted references · 67 canonical work pages

  1. [21]

    Language of Stains: Tokenization Enhances Multiplex Immunofluorescence and Histology Image Synthesis | bioRxiv

    “Language of Stains: Tokenization Enhances Multiplex Immunofluorescence and Histology Image Synthesis | bioRxiv.” Accessed: Apr. 22, 2025. [Online]. Available: https://www.biorxiv.org/content/10.1101/2025.03.04.641512v1

    work page doi:10.1101/2025.03.04.641512v1 2025

  2. [1]

    The human body at cellular resolution: the NIH Human Biomolecular Atlas Program,

    M. P . Snyder et al., “The human body at cellular resolution: the NIH Human Biomolecular Atlas Program,” Nature, vol. 574, no. 7777, pp. 187–192, Oct. 2019, doi: 10.1038/s41586-019-1629-x

    doi:10.1038/s41586-019-1629-x 2019

  3. [2]

    Tumour atlases enable researchers to navigate through cancers,

    “Tumour atlases enable researchers to navigate through cancers,” Nature, Oct. 2024, doi: 10.1038/d41586-024-03498-9

    work page doi:10.1038/d41586-024-03498-9 2024

  4. [3]

    Spatial mapping of cellular senescence: emerging challenges and opportunities,

    A. U. Gurkar et al., “Spatial mapping of cellular senescence: emerging challenges and opportunities,” Nat. Aging, vol. 3, no. 7, pp. 776–790, Jul. 2023, doi: 10.1038/s43587- 023-00446-6

    work page doi:10.1038/s43587- 2023

  5. [4]

    CROST: a comprehensive repository of spatial transcriptomics,

    G. Wang, S. Wu, Z. Xiong, H. Qu, X. Fang, and Y . Bao, “CROST: a comprehensive repository of spatial transcriptomics,” Nucleic Acids Res., vol. 52, no. D1, pp. D882– D890, Jan. 2024, doi: 10.1093/nar/gkad782

    work page doi:10.1093/nar/gkad782 2024

  6. [5]

    STOmicsDB: a comprehensive database for spatial transcriptomics data sharing, analysis and visualization,

    Z. Xu et al., “STOmicsDB: a comprehensive database for spatial transcriptomics data sharing, analysis and visualization,” Nucleic Acids Res., vol. 52, no. D1, pp. D1053– D1061, Jan. 2024, doi: 10.1093/nar/gkad933

    work page doi:10.1093/nar/gkad933 2024

  7. [6]

    RudolfV: A Foundation Model by Pathologists for Pathologists,

    J. Dippel et al., “RudolfV: A Foundation Model by Pathologists for Pathologists,” Jun. 11, 2024, arXiv: arXiv:2401.04079. doi: 10.48550/arXiv.2401.04079

    arXiv doi:10.48550/arxiv.2401.04079 2024

  8. [7]

    An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale,

    A. Dosovitskiy et al., “An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale,” Jun. 03, 2021, arXiv: arXiv:2010.11929. doi: 10.48550/arXiv.2010.11929

    arXiv doi:10.48550/arxiv.2010.11929 2021

Show all 135 references
  1. [8]

    DINOv2: Learning Robust Visual Features without Supervision,

    M. Oquab et al., “DINOv2: Learning Robust Visual Features without Supervision,” Feb. 02, 2024, arXiv: arXiv:2304.07193. doi: 10.48550/arXiv.2304.07193

    2024 doi arXiv

  2. [9]

    Virchow: A Million-Slide Digital Pathology Foundation Model,

    E. Vorontsov et al., “Virchow: A Million-Slide Digital Pathology Foundation Model,” Jan. 18, 2024, arXiv: arXiv:2309.07778. doi: 10.48550/arXiv.2309.07778. 17

    2024 doi arXiv

  3. [10]

    Towards a general-purpose foundation model for computational pathology,

    R. J. Chen et al., “Towards a general-purpose foundation model for computational pathology,” Nat. Med., vol. 30, no. 3, pp. 850–862, Mar. 2024, doi: 10.1038/s41591- 024-02857-3

    2024 doi

  4. [11]

    Towards A Generalizable Pathology Foundation Model via Unified Knowledge Distillation,

    J. Ma et al., “Towards A Generalizable Pathology Foundation Model via Unified Knowledge Distillation,” Aug. 03, 2024, arXiv: arXiv:2407.18449. doi: 10.48550/arXiv.2407.18449

    2024 doi arXiv

  5. [12]

    From whole-slide image to biomarker prediction: end-to- end weakly supervised deep learning in computational pathology,

    O. S. M. El Nahhas et al., “From whole-slide image to biomarker prediction: end-to- end weakly supervised deep learning in computational pathology,” Nat. Protoc., pp. 1–24, Sep. 2024, doi: 10.1038/s41596-024-01047-2

    2024 doi

  6. [13]

    Language Models are Few-Shot Learners,

    T. B. Brown et al., “Language Models are Few-Shot Learners,” Jul. 22, 2020, arXiv: arXiv:2005.14165. doi: 10.48550/arXiv.2005.14165

    2020 doi arXiv

  7. [14]

    BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding,

    J. Devlin, M.-W. Chang, K. Lee, and K. Toutanova, “BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding,” May 24, 2019, arXiv: arXiv:1810.04805. doi: 10.48550/arXiv.1810.04805

    2019 doi arXiv

  8. [15]

    LLaMA: Open and Efficient Foundation Language Models,

    H. Touvron et al., “LLaMA: Open and Efficient Foundation Language Models,” Feb. 27, 2023, arXiv: arXiv:2302.13971. doi: 10.48550/arXiv.2302.13971

    2023 doi arXiv

  9. [16]

    A visual-language foundation model for computational pathology,

    M. Y . Lu et al., “A visual-language foundation model for computational pathology,” Nat. Med., vol. 30, no. 3, pp. 863–874, Mar. 2024, doi: 10.1038/s41591-024-02856-4

    2024 doi

  10. [17]

    A multimodal generative AI copilot for human pathology,

    M. Y . Lu et al., “A multimodal generative AI copilot for human pathology,” Nature, vol. 634, no. 8033, pp. 466–473, Oct. 2024, doi: 10.1038/s41586-024-07618-3

    2024 doi

  11. [18]

    Multimodal Whole Slide Foundation Model for Pathology,

    T. Ding et al., “Multimodal Whole Slide Foundation Model for Pathology,” Nov. 29, 2024, arXiv: arXiv:2411.19666. doi: 10.48550/arXiv.2411.19666

    2024 doi arXiv

  12. [19]

    Deep learning in spatially resolved transcriptomics: a comprehensive technical view,

    R. Zahedi et al., “Deep learning in spatially resolved transcriptomics: a comprehensive technical view,” Brief. Bioinform., vol. 25, no. 2, p. bbae082, Mar. 2024, doi: 10.1093/bib/bbae082

    2024 doi

  13. [20]

    Transformers in single-cell omics: a review and new perspectives,

    A. Szałata et al., “Transformers in single-cell omics: a review and new perspectives,” Nat. Methods, vol. 21, no. 8, pp. 1430–1443, Aug. 2024, doi: 10.1038/s41592-024- 02353-z

    2024 doi

  14. [22]

    scGPT: toward building a foundation model for single-cell multi-omics using generative AI,

    H. Cui et al., “scGPT: toward building a foundation model for single-cell multi-omics using generative AI,” Nat. Methods, vol. 21, no. 8, pp. 1470–1480, Aug. 2024, doi: 10.1038/s41592-024-02201-0

    2024 doi

  15. [23]

    scBERT as a large-scale pretrained deep language model for cell type annotation of single-cell RNA-seq data,

    F . Yang et al., “scBERT as a large-scale pretrained deep language model for cell type annotation of single-cell RNA-seq data,” Nat. Mach. Intell., vol. 4, no. 10, pp. 852– 866, Oct. 2022, doi: 10.1038/s42256-022-00534-z

    2022 doi

  16. [24]

    Gene2vec: distributed representation of genes based on co-expression,

    J. Du, P . Jia, Y . Dai, C. Tao, Z. Zhao, and D. Zhi, “Gene2vec: distributed representation of genes based on co-expression,” BMC Genomics, vol. 20, no. 1, p. 82, Feb. 2019, doi: 10.1186/s12864-018-5370-x

    2019 doi

  17. [25]

    Large-scale foundation model on single-cell transcriptomics,

    M. Hao et al., “Large-scale foundation model on single-cell transcriptomics,” Nat. Methods, vol. 21, no. 8, pp. 1481–1491, Aug. 2024, doi: 10.1038/s41592-024-02305- 7. 18

    2024 doi

  18. [26]

    CellPLM: Pre-training of Cell Language Model Beyond Single Cells,

    H. Wen et al., “CellPLM: Pre-training of Cell Language Model Beyond Single Cells,” Oct. 05, 2023, bioRxiv. doi: 10.1101/2023.10.03.560734

    2023 doi

  19. [27]

    Generative pretraining from large-scale transcriptomes for single-cell deciphering,

    H. Shen et al., “Generative pretraining from large-scale transcriptomes for single-cell deciphering,” iScience, vol. 26, no. 5, p. 106536, May 2023, doi: 10.1016/j.isci.2023.106536

    2023

  20. [28]

    xTrimoGene: An Efficient and Scalable Representation Learner for Single-Cell RNA-Seq Data,

    J. Gong et al., “xTrimoGene: An Efficient and Scalable Representation Learner for Single-Cell RNA-Seq Data,” Feb. 24, 2024, arXiv: arXiv:2311.15156. doi: 10.48550/arXiv.2311.15156

    2024 doi arXiv

  21. [29]

    Transformer for one stop interpretable cell type annotation,

    J. Chen, H. Xu, W. Tao, Z. Chen, Y . Zhao, and J.-D. J. Han, “Transformer for one stop interpretable cell type annotation,” Nat. Commun., vol. 14, no. 1, p. 223, Jan. 2023, doi: 10.1038/s41467-023-35923-4

    2023 doi

  22. [30]

    Transfer learning enables predictions in network biology,

    C. V . Theodoris et al., “Transfer learning enables predictions in network biology,” Nature, vol. 618, no. 7965, pp. 616–624, Jun. 2023, doi: 10.1038/s41586-023-06139- 9

    2023 doi

  23. [31]

    Nicheformer: a foundation model for single-cell and spatial omics,

    A. C. Schaar et al., “Nicheformer: a foundation model for single-cell and spatial omics,” Apr. 17, 2024, bioRxiv. doi: 10.1101/2024.04.15.589472

    2024 doi

  24. [32]

    Single Cells Are Spatial Tokens: Transformers for Spatial Transcriptomic Data Imputation,

    H. Wen et al., “Single Cells Are Spatial Tokens: Transformers for Spatial Transcriptomic Data Imputation,” Feb. 16, 2024, arXiv: arXiv:2302.03038. Accessed: Sep. 11, 2024. [Online]. Available: http://arxiv.org/abs/2302.03038

    2024 arXiv

  25. [33]

    stEnTrans: Transformer-based deep learning for spatial transcriptomics enhancement,

    S. Xue, F . Zhu, C. Wang, and W. Min, “stEnTrans: Transformer-based deep learning for spatial transcriptomics enhancement,” Jul. 11, 2024, arXiv: arXiv:2407.08224. doi: 10.48550/arXiv.2407.08224

    2024 doi arXiv

  26. [34]

    Distinct spatial immune microlandscapes are independently associated with outcomes in triple-negative breast cancer,

    J. M. Carter et al., “Distinct spatial immune microlandscapes are independently associated with outcomes in triple-negative breast cancer,” Nat. Commun., vol. 14, no. 1, p. 2215, Apr . 2023, doi: 10.1038/s41467-023-37806-0

    2023 doi

  27. [35]

    Single-cell and spatial profiling identify three response trajectories to pembrolizumab and radiation therapy in triple negative breast cancer,

    S. L. Shiao et al., “Single-cell and spatial profiling identify three response trajectories to pembrolizumab and radiation therapy in triple negative breast cancer,” Cancer Cell, vol. 42, no. 1, pp. 70-84.e8, Jan. 2024, doi: 10.1016/j.ccell.2023.12.012

    2024 doi

  28. [36]

    Deep Profiling of Mouse Splenic Architecture with CODEX Multiplexed Imaging,

    Y . Goltsev et al., “Deep Profiling of Mouse Splenic Architecture with CODEX Multiplexed Imaging,” Cell, vol. 174, no. 4, pp. 968-981.e15, Aug. 2018, doi: 10.1016/j.cell.2018.07.010

    2018 doi

  29. [37]

    Highly multiplexed tissue imaging using repeated oligonucleotide exchange reaction,

    J. Kennedy-Darling et al., “Highly multiplexed tissue imaging using repeated oligonucleotide exchange reaction,” Eur. J. Immunol., vol. 51, no. 5, pp. 1262–1277, 2021, doi: 10.1002/eji.202048891

    2021 doi

  30. [38]

    Highly multiplexed imaging of tumor tissues with subcellular resolution by mass cytometry,

    C. Giesen et al., “Highly multiplexed imaging of tumor tissues with subcellular resolution by mass cytometry,” Nat. Methods, vol. 11, no. 4, pp. 417–422, Apr. 2014, doi: 10.1038/nmeth.2869

    2014 doi

  31. [39]

    IBEX: an iterative immunolabeling and chemical bleaching method for high-content imaging of diverse tissues,

    A. J. Radtke et al., “IBEX: an iterative immunolabeling and chemical bleaching method for high-content imaging of diverse tissues,” Nat. Protoc., vol. 17, no. 2, pp. 378–401, Feb. 2022, doi: 10.1038/s41596-021-00644-9

    2022 doi

  32. [40]

    Computational immune synapse analysis reveals T-cell interactions in distinct tumor microenvironments,

    V . G. Wang et al., “Computational immune synapse analysis reveals T-cell interactions in distinct tumor microenvironments,” Commun. Biol., vol. 7, no. 1, pp. 1–17, Sep. 2024, doi: 10.1038/s42003-024-06902-2. 19

    2024 doi

  33. [41]

    Prediction of Outcome from Spatial Protein Profiling of Triple- Negative Breast Cancers,

    A. F . Pour et al., “Prediction of Outcome from Spatial Protein Profiling of Triple- Negative Breast Cancers,” Apr. 18, 2025, bioRxiv. doi: 10.1101/2025.04.18.649541

    2025 doi

  34. [42]

    High-dimensional imaging using combinatorial channel multiplexing and deep learning,

    R. Ben-Uri et al., “High-dimensional imaging using combinatorial channel multiplexing and deep learning,” Nat. Biotechnol., pp. 1–14, Mar. 2025, doi: 10.1038/s41587-025- 02585-0

    2025 doi

  35. [43]

    Multi-view deep learning of highly multiplexed imaging data improves association of cell states with clinical outcomes,

    S. Ayub, H. W. Jackson, A. Selega, and K. R. Campbell, “Multi-view deep learning of highly multiplexed imaging data improves association of cell states with clinical outcomes,” Mar. 17, 2025, bioRxiv. doi: 10.1101/2025.03.14.643377

    2025 doi

  36. [44]

    A Foundation Model for Spatial Proteomics,

    M. Shaban et al., “A Foundation Model for Spatial Proteomics,” Jun. 05, 2025, arXiv: arXiv:2506.03373. doi: 10.48550/arXiv.2506.03373

    2025 doi arXiv

  37. [45]

    Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead,

    C. Rudin, “Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead,” Nat. Mach. Intell., vol. 1, no. 5, pp. 206–215, May 2019, doi: 10.1038/s42256-019-0048-x

    2019 doi

  38. [46]

    Towards Robust Interpretability with Self- Explaining Neural Networks,

    D. Alvarez-Melis and T. S. Jaakkola, “Towards Robust Interpretability with Self- Explaining Neural Networks,” Dec. 03, 2018, arXiv: arXiv:1806.07538. doi: 10.48550/arXiv.1806.07538

    2018 doi arXiv

  39. [47]

    Attention is not Explanation,

    S. Jain and B. C. Wallace, “Attention is not Explanation,” in Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), J. Burstein, C. Doran, and T . Solori...

    2019 doi

  40. [48]

    Is Attention Interpretable?,

    S. Serrano and N. A. Smith, “Is Attention Interpretable?,” in Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, A. Korhonen, D. Traum, and L. Màrquez, Eds., Florence, Italy: Association for Computational Linguistics, Jul. 2019, pp. 2931–2...

    2019 doi

  41. [49]

    Why Attentions May Not Be Interpretable?,

    B. Bai, J. Liang, G. Zhang, H. Li, K. Bai, and F . Wang, “Why Attentions May Not Be Interpretable?,” Jun. 03, 2021, arXiv: arXiv:2006.05656. doi: 10.48550/arXiv.2006.05656

    2021 doi arXiv

  42. [50]

    A Unified Approach to Interpreting Model Predictions,

    S. Lundberg and S.-I. Lee, “A Unified Approach to Interpreting Model Predictions,” Nov. 25, 2017, arXiv: arXiv:1705.07874. doi: 10.48550/arXiv.1705.07874

    2017 doi arXiv

  43. [51]

    ‘Why Should I Trust You?’: Explaining the Predictions of Any Classifier,

    M. T. Ribeiro, S. Singh, and C. Guestrin, “‘Why Should I Trust You?’: Explaining the Predictions of Any Classifier,” Aug. 09, 2016, arXiv: arXiv:1602.04938. doi: 10.48550/arXiv.1602.04938

    2016 doi arXiv

  44. [52]

    Context-guided Responsible Data Augmentation with Diffusion Models,

    K. Islam and N. Akhtar, “Context-guided Responsible Data Augmentation with Diffusion Models,” Mar. 12, 2025, arXiv: arXiv:2503.10687. doi: 10.48550/arXiv.2503.10687

    2025 doi arXiv

  45. [53]

    Thermodynamics-inspired explanations of artificial intelligence,

    S. Mehdi and P . Tiwary, “Thermodynamics-inspired explanations of artificial intelligence,” Nat. Commun., vol. 15, no. 1, p. 7859, Sep. 2024, doi: 10.1038/s41467- 024-51970-x

    2024 doi

  46. [54]

    Tissue registration and exploration user interfaces in support of a human reference atlas,

    K. Börner et al., “Tissue registration and exploration user interfaces in support of a human reference atlas,” Commun. Biol., vol. 5, no. 1, pp. 1–9, Dec. 2022, doi: 10.1038/s42003-022-03644-x. 20

    2022 doi

  47. [55]

    Advances and prospects for the Human BioMolecular Atlas Program (HuBMAP),

    S. Jain et al., “Advances and prospects for the Human BioMolecular Atlas Program (HuBMAP),” Nat. Cell Biol., vol. 25, no. 8, pp. 1089–1100, Aug. 2023, doi: 10.1038/s41556-023-01194-w

    2023 doi

  48. [56]

    Graph Fourier transform for spatial omics representation and analyses of complex organs,

    Y . Chang et al., “Graph Fourier transform for spatial omics representation and analyses of complex organs,” Res. Sq., p. rs.3.rs-3952048, Feb. 2024, doi: 10.21203/rs.3.rs-3952048/v1

    2024 doi

  49. [57]

    Imaging mass cytometry and multiplatform genomics define the phenogenomic landscape of breast cancer,

    H. R. Ali et al., “Imaging mass cytometry and multiplatform genomics define the phenogenomic landscape of breast cancer,” Nat. Cancer, vol. 1, no. 2, pp. 163–175, Feb. 2020, doi: 10.1038/s43018-020-0026-6

    2020 doi

  50. [58]

    The single-cell pathology landscape of breast cancer,

    H. W. Jackson et al., “The single-cell pathology landscape of breast cancer,” Nature, vol. 578, no. 7796, pp. 615–620, Feb. 2020, doi: 10.1038/s41586-019-1876-x

    2020 doi

  51. [59]

    Probing the Limits to Positional Information,

    T. Gregor, D. W. Tank, E. F . Wieschaus, and W. Bialek, “Probing the Limits to Positional Information,” Cell, vol. 130, no. 1, pp. 153–164, Jul. 2007, doi: 10.1016/j.cell.2007.05.025

    2007 doi

  52. [60]

    Decoding cellular communication: An information theoretic perspective on cytokine and endocrine signaling,

    F . Schaper, T. Jetka, and A. Dittrich, “Decoding cellular communication: An information theoretic perspective on cytokine and endocrine signaling,” Curr. Opin. Endocr. Metab. Res., vol. 24, p. 100351, Jun. 2022, doi: 10.1016/j.coemr.2022.100351

    2022

  53. [61]

    Concepts and Applications of Information Theory to Immuno-Oncology,

    A. Karolak, S. Branciamore, J. S. McCune, P . P . Lee, A. S. Rodin, and R. C. Rockne, “Concepts and Applications of Information Theory to Immuno-Oncology,” Trends Cancer, vol. 7, no. 4, pp. 335–346, Apr. 2021, doi: 10.1016/j.trecan.2020.12.013

    2021 doi

  54. [62]

    Integrative spatial analysis reveals a multi-layered organization of glioblastoma,

    A. C. Greenwald et al., “Integrative spatial analysis reveals a multi-layered organization of glioblastoma,” Cell, vol. 187, no. 10, pp. 2485-2501.e26, May 2024, doi: 10.1016/j.cell.2024.03.029

    2024 doi

  55. [63]

    The information bottleneck method,

    N. Tishby, F . C. Pereira, and W. Bialek, “The information bottleneck method,” Apr. 24, 2000, arXiv: arXiv:physics/0004057. doi: 10.48550/arXiv.physics/0004057

    2000 doi arXiv

  56. [64]

    Identifying maximally informative signal- aware representations of single-cell data using the Information Bottleneck,

    S. Dubnov, Z. Piran, H. Soreq, and M. Nitzan, “Identifying maximally informative signal- aware representations of single-cell data using the Information Bottleneck,” Jul. 17, 2024, bioRxiv. doi: 10.1101/2024.05.22.595292

    2024 doi

  57. [65]

    Deep Learning and the Information Bottleneck Principle,

    N. Tishby and N. Zaslavsky, “Deep Learning and the Information Bottleneck Principle,” Mar. 09, 2015, arXiv: arXiv:1503.02406. doi: 10.48550/arXiv.1503.02406

    2015 doi arXiv

  58. [66]

    An Information Theoretic Interpretation to Deep Neural Networks,

    X. Xu, S.-L. Huang, L. Zheng, and G. W. Wornell, “An Information Theoretic Interpretation to Deep Neural Networks,” Entropy, vol. 24, no. 1, p. 135, Jan. 2022, doi: 10.3390/e24010135

    2022 doi

  59. [67]

    DALL·E 3

    “DALL·E 3.” Accessed: Feb. 14, 2025. [Online]. Available: https://openai.com/index/dall-e-3/

    2025

  60. [68]

    Stability AI Image Models,

    “Stability AI Image Models,” Stability AI. Accessed: Feb. 14, 2025. [Online]. Available: https://stability.ai/stable-image

    2025

  61. [69]

    Score- Based Generative Modeling through Stochastic Differential Equations,

    Y . Song, J. Sohl-Dickstein, D. P . Kingma, A. Kumar, S. Ermon, and B. Poole, “Score- Based Generative Modeling through Stochastic Differential Equations,” Feb. 10, 2021, arXiv: arXiv:2011.13456. doi: 10.48550/arXiv.2011.13456

    2021 doi arXiv

  62. [70]

    High-Resolution Image Synthesis with Latent Diffusion Models,

    R. Rombach, A. Blattmann, D. Lorenz, P . Esser, and B. Ommer, “High-Resolution Image Synthesis with Latent Diffusion Models,” Apr. 13, 2022, arXiv: arXiv:2112.10752. doi: 10.48550/arXiv.2112.10752. 21

    2022 doi arXiv

  63. [71]

    Scalable Diffusion Models with Transformers,

    W. Peebles and S. Xie, “Scalable Diffusion Models with Transformers,” arXiv.org. Accessed: Oct. 02, 2024. [Online]. Available: https://arxiv.org/abs/2212.09748v2

    2024 arXiv

  64. [72]

    stDiff: a diffusion model for imputing spatial transcriptomics through single-cell transcriptomics | Briefings in Bioinformatics | Oxford Academic

    “stDiff: a diffusion model for imputing spatial transcriptomics through single-cell transcriptomics | Briefings in Bioinformatics | Oxford Academic.” Accessed: Oct. 02,

  65. [73]

    SpaDiT: Diffusion Transformer for Spatial Gene Expression Prediction using scRNA-seq,

    X. Li, F . Zhu, and W. Min, “SpaDiT: Diffusion Transformer for Spatial Gene Expression Prediction using scRNA-seq,” Jul. 18, 2024, arXiv: arXiv:2407.13182. doi: 10.48550/arXiv.2407.13182

    2024 doi arXiv

  66. [74]

    DiffuST: a latent diffusion model for spatial transcriptomics denoising,

    S. Jiao et al., “DiffuST: a latent diffusion model for spatial transcriptomics denoising,” Jun. 23, 2024, bioRxiv. doi: 10.1101/2024.06.19.599672

    2024 doi

  67. [75]

    SpatialDiffusion: Predicting Spatial Transcriptomics with Denoising Diffusion Probabilistic Models,

    S. A. Khan, V . Lagani, R. Lehmann, N. A. Kiani, D. Gomez-Cabrero, and J. Tegner, “SpatialDiffusion: Predicting Spatial Transcriptomics with Denoising Diffusion Probabilistic Models,” May 21, 2024, bioRxiv. doi: 10.1101/2024.05.21.595094

    2024 doi

  68. [76]

    stMCDI: Masked Conditional Diffusion Model with Graph Neural Network for Spatial Transcriptomics Data Imputation,

    X. Li, W. Min, S. Wang, C. Wang, and T. Xu, “stMCDI: Masked Conditional Diffusion Model with Graph Neural Network for Spatial Transcriptomics Data Imputation,” Mar. 16, 2024, arXiv: arXiv:2403.10863. doi: 10.48550/arXiv.2403.10863

    2024 doi arXiv

  69. [77]

    The statistical thermodynamics of generative diffusion models: Phase transitions, symmetry breaking and critical instability,

    L. Ambrogioni, “The statistical thermodynamics of generative diffusion models: Phase transitions, symmetry breaking and critical instability,” arXiv.org. Accessed: Oct. 02,

  70. [78]

    Generative diffusion in very large dimensions,

    G. Biroli and M. Mézard, “Generative diffusion in very large dimensions,” J. Stat. Mech. Theory Exp., vol. 2023, no. 9, p. 093402, Sep. 2023, doi: 10.1088/1742-5468/acf8ba

    2023 doi

  71. [79]

    Available: https://arxiv.org/abs/2310.17467v4

    [Online]. Available: https://arxiv.org/abs/2310.17467v4

    arXiv

  72. [80]

    Mapping the topography of spatial gene expression with interpretable deep learning,

    U. Chitra et al., “Mapping the topography of spatial gene expression with interpretable deep learning,” BioRxiv Prepr. Serv. Biol., p. 2023.10.10.561757, Oct. 2023, doi: 10.1101/2023.10.10.561757

    2023 doi

  73. [81]

    A phase transition in diffusion models reveals the hierarchical nature of data,

    A. Sclocchi, A. Favero, and M. Wyart, “A phase transition in diffusion models reveals the hierarchical nature of data,” Proc. Natl. Acad. Sci., vol. 122, no. 1, p. e2408799121, Jan. 2025, doi: 10.1073/pnas.2408799121

    2025 doi

  74. [82]

    A mathematical model for predicting the spatiotemporal response of breast cancer cells treated with doxorubicin,

    H. J. M. Miniere et al., “A mathematical model for predicting the spatiotemporal response of breast cancer cells treated with doxorubicin,” Cancer Biol. Ther., vol. 25, no. 1, p. 2321769, Dec. 2024, doi: 10.1080/15384047.2024.2321769

    2024

  75. [83]

    A Reaction-Diffusion Model of Cancer Invasion,

    R. A. Gatenby and E. T. Gawlinski, “A Reaction-Diffusion Model of Cancer Invasion,” Cancer Res., vol. 56, no. 24, pp. 5745–5753, Dec. 1996

    1996

  76. [84]

    Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations,

    M. Raissi, P . Perdikaris, and G. E. Karniadakis, “Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations,” J. Comput. Phys., vol. 378, pp. 686–707, Feb. 2019, doi: 10.1016/j.jcp.2...

    2019 doi

  77. [85]

    A mechanically coupled reaction–diffusion model that incorporates intra-tumoural heterogeneity to predict in vivo glioma growth,

    D. A. Hormuth et al., “A mechanically coupled reaction–diffusion model that incorporates intra-tumoural heterogeneity to predict in vivo glioma growth,” J. R. Soc. Interface, vol. 14, no. 128, p. 20161010, Mar. 2017, doi: 10.1098/rsif.2016.1010

    2017

  78. [86]

    NeuroVelo: interpretable learning of temporal cellular dynamics from single-cell data,

    I. K. Boudjelthia, S. Milite, N. E. Kazwini, Y . Huang, A. Sottoriva, and G. Sanguinetti, “NeuroVelo: interpretable learning of temporal cellular dynamics from single-cell data,” Jun. 10, 2024, bioRxiv. doi: 10.1101/2023.11.17.567500

    2024 doi

  79. [87]

    A physics-informed neural SDE network for learning cellular dynamics from time-series scRNA-seq data,

    Q. Jiang and L. Wan, “A physics-informed neural SDE network for learning cellular dynamics from time-series scRNA-seq data,” Bioinformatics, vol. 40, no. Supplement_2, pp. ii120–ii127, Sep. 2024, doi: 10.1093/bioinformatics/btae400. 22

    2024 doi

  80. [88]

    SpaCCC: Large language model-based cell-cell communication inference for spatially resolved transcriptomic data,

    B. Ji, L. Xu, and S. Peng, “SpaCCC: Large language model-based cell-cell communication inference for spatially resolved transcriptomic data,” Feb. 23, 2024, bioRxiv. doi: 10.1101/2024.02.21.581369

    2024 doi

  81. [89]

    Decoding functional cell–cell communication events by multi-view graph learning on spatial transcriptomics,

    H. Li et al., “Decoding functional cell–cell communication events by multi-view graph learning on spatial transcriptomics,” Brief. Bioinform., vol. 24, no. 6, p. bbad359, Nov. 2023, doi: 10.1093/bib/bbad359

    2023 doi

  82. [90]

    How to Learn More? Exploring Kolmogorov-Arnold Networks for Hyperspectral Image Classification,

    A. Jamali, S. K. Roy, D. Hong, B. Lu, and P . Ghamisi, “How to Learn More? Exploring Kolmogorov-Arnold Networks for Hyperspectral Image Classification,” Jun. 22, 2024, arXiv: arXiv:2406.15719. doi: 10.48550/arXiv.2406.15719

    2024 doi arXiv

  83. [91]

    KAN: Kolmogorov–Arnold Networks

    “KAN: Kolmogorov–Arnold Networks.” Accessed: Oct. 11, 2024. [Online]. Available: https://arxiv.org/html/2404.19756v1

    2024 arXiv

  84. [92]

    Mechanical properties of human tumour tissues and their implications for cancer development,

    A. Massey et al., “Mechanical properties of human tumour tissues and their implications for cancer development,” Nat. Rev. Phys., vol. 6, no. 4, pp. 269–282, Apr. 2024, doi: 10.1038/s42254-024-00707-2

    2024 doi

  85. [93]

    sepal: identifying transcript profiles with spatial patterns by diffusion-based modeling,

    A. Andersson and J. Lundeberg, “sepal: identifying transcript profiles with spatial patterns by diffusion-based modeling,” Bioinformatics, vol. 37, no. 17, pp. 2644– 2650, Sep. 2021, doi: 10.1093/bioinformatics/btab164

    2021 doi

  86. [94]

    Cell adhesion in cancer: Beyond the migration of single cells,

    M. Janiszewska, M. C. Primi, and T. Izard, “Cell adhesion in cancer: Beyond the migration of single cells,” J. Biol. Chem., vol. 295, no. 8, pp. 2495–2505, Feb. 2020, doi: 10.1074/jbc.REV119.007759

    2020 doi

  87. [95]

    Modulating extracellular matrix stiffness: a strategic approach to boost cancer immunotherapy,

    Z. Mai, Y . Lin, P . Lin, X. Zhao, and L. Cui, “Modulating extracellular matrix stiffness: a strategic approach to boost cancer immunotherapy,” Cell Death Dis., vol. 15, no. 5, pp. 1–16, May 2024, doi: 10.1038/s41419-024-06697-4

    2024 doi

  88. [96]

    Mechanical Forces in Tumor Angiogenesis,

    M. R. Zanotelli and C. A. Reinhart-King, “Mechanical Forces in Tumor Angiogenesis,” Adv. Exp. Med. Biol., vol. 1092, pp. 91–112, 2018, doi: 10.1007/978-3-319-95294-9_6

    2018 doi

  89. [97]

    The physics of cancer: the role of physical interactions and mechanical forces in metastasis,

    D. Wirtz, K. Konstantopoulos, and P . C. Searson, “The physics of cancer: the role of physical interactions and mechanical forces in metastasis,” Nat. Rev. Cancer, vol. 11, no. 7, pp. 512–522, Jun. 2011, doi: 10.1038/nrc3080

    2011 doi

  90. [98]

    Advances in cancer mechanobiology: Metastasis, mechanics, and materials,

    A. J. Clevenger, M. K. McFarlin, J. P . M. Gorley, S. C. Solberg, A. K. Madyastha, and S. A. Raghavan, “Advances in cancer mechanobiology: Metastasis, mechanics, and materials,” APL Bioeng., vol. 8, no. 1, p. 011502, Mar. 2024, doi: 10.1063/5.0186042

    2024 doi

  91. [99]

    A computational pipeline for spatial mechano-transcriptomics,

    A. Hallou, R. He, B. D. Simons, and B. Dumitrascu, “A computational pipeline for spatial mechano-transcriptomics,” Aug. 05, 2023, bioRxiv. doi: 10.1101/2023.08.03.551894

    2023 doi

  92. [100]

    Physics-informed neural network estimation of material properties in soft tissue nonlinear biomechanical models,

    F . Caforio et al., “Physics-informed neural network estimation of material properties in soft tissue nonlinear biomechanical models,” Comput. Mech., Jul. 2024, doi: 10.1007/s00466-024-02516-x. 23

    2024 doi

  93. [101]

    Mechanobiology of collective cell behaviours,

    B. Ladoux and R.-M. Mège, “Mechanobiology of collective cell behaviours,” Nat. Rev. Mol. Cell Biol., vol. 18, no. 12, pp. 743–757, Dec. 2017, doi: 10.1038/nrm.2017.98

    2017 doi

  94. [102]

    Tumour evolution and microenvironment interactions in 2D and 3D space,

    C.-K. Mo et al., “Tumour evolution and microenvironment interactions in 2D and 3D space,” Nature, vol. 634, no. 8036, pp. 1178–1186, Oct. 2024, doi: 10.1038/s41586- 024-08087-4

    2024 doi

  95. [103]

    Haplotype-aware analysis of somatic copy number variations from single-cell transcriptomes,

    T. Gao et al., “Haplotype-aware analysis of somatic copy number variations from single-cell transcriptomes,” Nat. Biotechnol., vol. 41, no. 3, pp. 417–426, Mar. 2023, doi: 10.1038/s41587-022-01468-y

    2023 doi

  96. [104]

    Spatiotemporal profiling defines persistence and resistance dynamics during targeted treatment of melanoma,

    J. C. Rubinstein et al., “Spatiotemporal profiling defines persistence and resistance dynamics during targeted treatment of melanoma,” Feb. 05, 2024, bioRxiv. doi: 10.1101/2024.02.02.577085

    2024 doi

  97. [105]

    Single-cell RNA-seq highlights intratumoral heterogeneity in primary glioblastoma,

    A. P . Patel et al., “Single-cell RNA-seq highlights intratumoral heterogeneity in primary glioblastoma,” Science, vol. 344, no. 6190, pp. 1396–1401, Jun. 2014, doi: 10.1126/science.1254257

    2014 doi

  98. [106]

    Inferring allele-specific copy number aberrations and tumor phylogeography from spatially resolved transcriptomics,

    C. Ma et al., “Inferring allele-specific copy number aberrations and tumor phylogeography from spatially resolved transcriptomics,” Nat. Methods, pp. 1–9, Oct. 2024, doi: 10.1038/s41592-024-02438-9

    2024 doi

  99. [107]

    CoT: a transformer-based method for inferring tumor clonal copy number substructure from scDNA-seq data,

    F . Liu, F . Shi, F . Du, X. Cao, and Z. Yu, “CoT: a transformer-based method for inferring tumor clonal copy number substructure from scDNA-seq data,” Brief. Bioinform., vol. 25, no. 3, p. bbae187, May 2024, doi: 10.1093/bib/bbae187

    2024 doi

  100. [108]

    Pan-cancer computational histopathology reveals mutations, tumor composition and prognosis,

    Y . Fu et al., “Pan-cancer computational histopathology reveals mutations, tumor composition and prognosis,” Nat. Cancer, vol. 1, no. 8, pp. 800–810, Aug. 2020, doi: 10.1038/s43018-020-0085-8

    2020 doi

  101. [109]

    Using Physics-Informed Neural Networks (PINNs) for Tumor Cell Growth Modeling,

    J. A. Rodrigues, “Using Physics-Informed Neural Networks (PINNs) for Tumor Cell Growth Modeling,” Mathematics, vol. 12, no. 8, Art. no. 8, Jan. 2024, doi: 10.3390/math12081195

    2024 doi

  102. [110]

    STalign: Alignment of spatial transcriptomics data using diffeomorphic metric mapping,

    K. Clifton et al., “STalign: Alignment of spatial transcriptomics data using diffeomorphic metric mapping,” Nat. Commun., vol. 14, no. 1, p. 8123, Dec. 2023, doi: 10.1038/s41467-023-43915-7

    2023 doi

  103. [111]

    Deep learning trained on hematoxylin and eosin tumor region of Interest predicts HER2 status and trastuzumab treatment response in HER2+ breast cancer,

    S. Farahmand et al., “Deep learning trained on hematoxylin and eosin tumor region of Interest predicts HER2 status and trastuzumab treatment response in HER2+ breast cancer,” Mod. Pathol., vol. 35, no. 1, pp. 44–51, Jan. 2022, doi: 10.1038/s41379-021- 00911-w

    2022 doi

  104. [112]

    Bayesian approach to single-cell differential expression analysis,

    P . V . Kharchenko, L. Silberstein, and D. T. Scadden, “Bayesian approach to single-cell differential expression analysis,” Nat. Methods, vol. 11, no. 7, pp. 740–742, Jul. 2014, doi: 10.1038/nmeth.2967

    2014 doi

  105. [113]

    Quantifying and correcting slide-to-slide variation in multiplexed immunofluorescence images,

    C. R. Harris et al., “Quantifying and correcting slide-to-slide variation in multiplexed immunofluorescence images,” Bioinformatics, vol. 38, no. 6, pp. 1700–1707, Mar. 2022, doi: 10.1093/bioinformatics/btab877

    2022 doi

  106. [114]

    TIST: Transcriptome and Histopathological Image Integrative Analysis for Spatial Transcriptomics,

    Y . Shan et al., “TIST: Transcriptome and Histopathological Image Integrative Analysis for Spatial Transcriptomics,” Genomics Proteomics Bioinformatics, vol. 20, no. 5, pp. 974–988, Oct. 2022, doi: 10.1016/j.gpb.2022.11.012. 24

    2022 doi

  107. [115]

    Integrating spatial gene expression and breast tumour morphology via deep learning,

    B. He et al., “Integrating spatial gene expression and breast tumour morphology via deep learning,” Nat. Biomed. Eng., vol. 4, no. 8, pp. 827–834, Aug. 2020, doi: 10.1038/s41551-020-0578-x

    2020 doi

  108. [116]

    Deciphering spatial domains from spatial multi-omics with SpatialGlue,

    Y . Long et al., “Deciphering spatial domains from spatial multi-omics with SpatialGlue,” Nat. Methods, vol. 21, no. 9, pp. 1658–1667, Sep. 2024, doi: 10.1038/s41592-024-02316-4

    2024 doi

  109. [117]

    Nextflow pipeline for Visium and H&E data from patient-derived xenograft samples,

    S. Domanskyi et al., “Nextflow pipeline for Visium and H&E data from patient-derived xenograft samples,” Cell Rep. Methods, vol. 4, no. 5, p. 100759, May 2024, doi: 10.1016/j.crmeth.2024.100759

    2024

  110. [118]

    The covariance environment defines cellular niches for spatial inference,

    D. Haviv et al., “The covariance environment defines cellular niches for spatial inference,” Nat. Biotechnol., vol. 43, no. 2, pp. 269–280, Feb. 2025, doi: 10.1038/s41587-024-02193-4

    2025 doi

  111. [119]

    Single-Cell Multimodal Prediction via Transformers,

    W. Tang et al., “Single-Cell Multimodal Prediction via Transformers,” in Proceedings of the 32nd ACM International Conference on Information and Knowledge Management, Oct. 2023, pp. 2422–2431. doi: 10.1145/3583780.3615061

    2023

  112. [120]

    A visual–omics foundation model to bridge histopathology image with transcriptomics,

    G. Wang et al., “A visual–omics foundation model to bridge histopathology image with transcriptomics,” Apr. 16, 2025, Research Square. doi: 10.21203/rs.3.rs- 5183775/v1

    2025 doi

  113. [121]

    Molecular-driven Foundation Model for Oncologic Pathology,

    A. Vaidya et al., “Molecular-driven Foundation Model for Oncologic Pathology,” Jan. 28, 2025, arXiv: arXiv:2501.16652. doi: 10.48550/arXiv.2501.16652

    2025 doi arXiv

  114. [122]

    MOSAIC is the world’s largest spatial multiomics dataset in oncology

    “MOSAIC is the world’s largest spatial multiomics dataset in oncology.” Accessed: Feb. 14, 2025. [Online]. Available: https://www.mosaic-research.com/

    2025

  115. [123]

    SPIN-AI: A Deep Learning Model That Identifies Spatially Predictive Genes,

    K. Meng-Lin et al., “SPIN-AI: A Deep Learning Model That Identifies Spatially Predictive Genes,” Biomolecules, vol. 13, no. 6, p. 895, May 2023, doi: 10.3390/biom13060895

    2023 doi

  116. [124]

    Collaborative Cross and Diversity Outbred data resources in the Mouse Phenome Database,

    M. A. Bogue, G. A. Churchill, and E. J. Chesler, “Collaborative Cross and Diversity Outbred data resources in the Mouse Phenome Database,” Mamm. Genome, vol. 26, no. 9, pp. 511–520, Oct. 2015, doi: 10.1007/s00335-015-9595-6

    2015 doi

  117. [125]

    Comprehensive single-cell aging atlas of healthy mammary tissues reveals shared epigenomic and transcriptomic signatures of aging and cancer,

    B. L. Angarola et al., “Comprehensive single-cell aging atlas of healthy mammary tissues reveals shared epigenomic and transcriptomic signatures of aging and cancer,” Nat. Aging, vol. 5, no. 1, pp. 122–143, Jan. 2025, doi: 10.1038/s43587-024- 00751-8

    2025 doi

  118. [126]

    Unraveling the genetics of arsenic toxicity with cellular morphology QTL,

    C. O’Connor et al., “Unraveling the genetics of arsenic toxicity with cellular morphology QTL,” PLOS Genet., vol. 20, no. 4, p. e1011248, Apr. 2024, doi: 10.1371/journal.pgen.1011248

    2024 doi

  119. [127]

    Mapping the genetic landscape establishing a tumor immune microenvironment favorable for anti-PD-1 response in mice and humans,

    D. A. Skelly et al., “Mapping the genetic landscape establishing a tumor immune microenvironment favorable for anti-PD-1 response in mice and humans,” Jul. 16, 2024, bioRxiv. doi: 10.1101/2024.07.11.603136

    2024 doi

  120. [128]

    Whole brain alignment of spatial transcriptomics between humans and mice with BrainAlign,

    B. Zhang, S. Zhang, and S. Zhang, “Whole brain alignment of spatial transcriptomics between humans and mice with BrainAlign,” Nat. Commun., vol. 15, no. 1, p. 6302, Jul. 2024, doi: 10.1038/s41467-024-50608-2

    2024 doi

  121. [129]

    Cross-species cell-type assignment from single-cell RNA-seq data by a heterogeneous graph neural network,

    X. Liu, Q. Shen, and S. Zhang, “Cross-species cell-type assignment from single-cell RNA-seq data by a heterogeneous graph neural network,” Genome Res., vol. 33, no. 1, pp. 96–111, Jan. 2023, doi: 10.1101/gr.276868.122

    2023 doi

  122. [130]

    Unsupervised discovery of tissue architecture in multiplexed imaging,

    J. Kim et al., “Unsupervised discovery of tissue architecture in multiplexed imaging,” Nat. Methods, vol. 19, no. 12, pp. 1653–1661, Dec. 2022, doi: 10.1038/s41592-022- 01657-2

    2022 doi

  123. [131]

    SAMPLER: unsupervised representations for rapid analysis of whole slide tissue images,

    P . Mukashyaka, T. B. Sheridan, A. Foroughi pour, and J. H. Chuang, “SAMPLER: unsupervised representations for rapid analysis of whole slide tissue images,” eBioMedicine, vol. 99, p. 104908, Jan. 2024, doi: 10.1016/j.ebiom.2023.104908. 25

    2024

  124. [132]

    Graph deep learning for the characterization of tumour microenvironments from spatial protein profiles in tissue specimens,

    Z. Wu et al., “Graph deep learning for the characterization of tumour microenvironments from spatial protein profiles in tissue specimens,” Nat. Biomed. Eng., vol. 6, no. 12, pp. 1435–1448, Dec. 2022, doi: 10.1038/s41551-022-00951-w

    2022 doi

  125. [133]

    Sampling and ranking spatial transcriptomics data embeddings to identify tissue architecture,

    Y . Lin et al., “Sampling and ranking spatial transcriptomics data embeddings to identify tissue architecture,” Front. Genet., vol. 13, Aug. 2022, doi: 10.3389/fgene.2022.912813

    2022

  126. [135]

    Deciphering cell–cell communication at single-cell resolution for spatial transcriptomics with subgraph-based graph attention network,

    W. Yang et al., “Deciphering cell–cell communication at single-cell resolution for spatial transcriptomics with subgraph-based graph attention network,” Nat. Commun., vol. 15, no. 1, p. 7101, Aug. 2024, doi: 10.1038/s41467-024-51329-2

    2024 doi

  127. [2024]

    Available: https://academic.oup.com/bib/article/25/3/bbae171/7646375

    [Online]. Available: https://academic.oup.com/bib/article/25/3/bbae171/7646375

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.