Pith. sign in

REVIEW 4 major objections 5 minor 24 references

Adaptive Spatial Transcriptomics Interpolation via Cross-modal Cross-slice Modeling

T0 review · 4 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 Two adjacent gene-expression tissue maps and their paired stained images suffice to synthesize any missing slice in between.

desk verdict Genuinely new ST interpolation task, but the arbitrary-position mechanism is under-specified in a load-bearing way and the single-dataset evidence needs error bars. read the letter →

arxiv 2505.10729 v1 pith:I627W7LV submitted 2025-05-15 eess.IV cs.CVq-bio.QM

NingFeng Que , Xiaofei Wang , Jingjing Chen , Yixuan Jiang , Chao Li This is my paper · ORCID
classification eess.IVcs.CVq-bio.QM
keywords spatialtranscriptomicssliceinterpolationcross-modalalignmentgeneco-expressiongraphdeformableconvolutionH&E-stainedimagesarbitrary-positiondeeplearning
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 proposes C2-STi, a deep network that takes two adjacent spatial transcriptomics (ST) slices and their paired hematoxylin-and-eosin (H&E) images and generates the ST slices that would lie at any requested intermediate position. The authors claim this is the first method to interpolate an arbitrary number of missing sections between adjacent ST slices, rather than only a single middle slice. The network couples three mechanisms: a cross-modal alignment gate that lets ST features borrow cellular detail from H&E, a pyramid graph that tracks co-expressed genes across scales, and a distance-aware modulation that turns a slice-position scalar into deformable-convolution adjustments. On the public HNSCC dataset they report consistently higher PSNR, SSIM, and Pearson correlation and lower RMSE than the tested frame-interpolation, medical-interpolation, and U-Net baselines in both single-slice and multi-slice settings. If the claim holds, a laboratory could produce dense three-dimensional gene-expression volumes from a few physically sectioned and stained slices.

What carries the argument

Distance-aware local structural modulation (DLSM) is the mechanism that makes arbitrary-position interpolation possible. It expresses each requested slice as a scalar $p_i = (i/s)\cdot\mathrm{Norm}(1+\alpha|\nabla H_i|)$, where the Sobel gradient of the H&E image reweights the plain slice index; that scalar is then embedded and split into a channel-adaptation branch and a spatial-adaptation branch whose outputs predict kernels, offsets, and masks for a deformable convolution. The deformable convolution reshapes the coarse forward and backward deformation features into features at position $p_i$, so one network can generate any intermediate slice rather than only the midpoint. The cross-modal alignment gate and the multi-gene co-expression graph are supporting modules: the former filters H&E features with gated attention before they touch ST features, and the latter defines a graph whose adjacency is Pearson correlation between gene-expression profiles and applies a graph convolution at each pyramid level.

What would settle it

Hold out real middle sections from HNSCC triples, train the model once with the proposed gradient-weighted position and once with alpha=0, and compare per-gene PSNR and Pearson correlation; if the gradient-weighted version is not consistently better on slice pairs with large H&E edge magnitudes, the distance model's core assumption fails.

Watch

Extended reading notes

Core claim

C2-STi is presented as the first algorithm to interpolate missing spatial transcriptomics (ST) slices at arbitrary intermediate positions between two adjacent ST slices. Given the two slices I0 and I1 and their paired H&E images H0 and H1, it outputs any requested number s of imputed slices {Iout^i}. Three modules carry the work: a cross-modal alignment gate that refines ST features with H&E-derived cellular structure; a pyramid gene co-expression correlation module that builds a Pearson-correlation adjacency matrix over genes and propagates it through a graph-convolution layer at three scales; and a distance-aware local structural modulation module that embeds a position scalar $p_i=(i/s)\mathrm{Norm}(1+\alpha|\nabla H_i|)$ into channel and spatial branches, producing kernels, offsets, and masks for deformable convolution. The network is trained with an L1 similarity loss against the ground-truth slice plus a gradient smoothness regularizer. On the HNSCC dataset the reported single-slice results beat the compared methods by at least 6.0% PSNR, 6.6% SSIM, 29.0% PCC, and 21.4% RMSE, with consistent improvements across 2-, 3-, and 4-slice interpolation, and the ablations show all three modules contribute.

Load-bearing premise

The method's central premise is that one scalar per intended slice, the slice index scaled by the H&E image's edge strength, correctly encodes where the missing tissue section lies between the two measured slices; if that scalar does not track how the tissue actually changes between sections, the synthesized intermediate slice is not biologically valid.

Editorial extensions

If this is right

  • A laboratory with only two successfully collected adjacent ST slices could reconstruct any number of intermediate expression maps in a single forward pass, replacing repeated high-cost sectioning for 3D analysis.
  • Output slices can be requested at arbitrary fractional positions, so a user can choose the density of the reconstructed volume after the model is trained.
  • The ablations show that removing the H&E cross-modal alignment hurts most, indicating that histology information is the largest contributor to the reported gains.
  • Interpolation quality is reported with standard image metrics, so downstream quality control could screen synthesized slices before using them in analysis.

Reading between the lines

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

  • A testable extension the paper does not run is applying the same position-scalar idea to temporal ST experiments, treating time between two observations as the interpolation coordinate; the gradient weight would then need to be a temporal-change proxy rather than an H&E edge map.
  • If the distance model works as claimed, interpolation error on a held-out middle section should correlate with the H&E gradient magnitude: sections with stronger edge structure should require larger learned offsets, a prediction one could check per region.
  • Since the paper evaluates on a single cancer dataset, the natural next question is cross-dataset transfer to a different tissue type; the authors do not claim such transfer.
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

4 major / 5 minor

Summary. The manuscript proposes C2-STi, a deep learning method for interpolating missing spatial transcriptomics (ST) slices between two adjacent ST slices, using the paired hematoxylin and eosin (H&E) images as additional input. The method combines a cross-modal alignment module, a pyramid gene co-expression correlation module, and a distance-aware local structural modulation (DLSM) module intended to enable interpolation at arbitrary intermediate positions. On the HNSCC dataset, the authors report higher PSNR, SSIM, and PCC and lower RMSE than five baselines for single-slice interpolation and for 2-, 3-, and 4-slice interpolation, and they provide ablation studies for each of the three proposed modules.

Significance. If the claims are correct, C2-STi would be a practically valuable contribution to spatial transcriptomics analysis, potentially reducing the cost of dense slicing by generating missing ST slices from adjacent pairs. The consistent margin over five baselines, the availability of code, and the ablation experiments are notable strengths, and I found no circularity: the model is trained on training patches and evaluated on held-out test patches against external baselines. However, the evidence base is a single dataset with a single split and no error bars or significance tests, and the arbitrary-position mechanism is under-specified. The significance of the claimed contribution is therefore conditional on resolving the methodological issues described below.

major comments (4)
  1. [Section 2.3, Cross-section distance modeling] The position encoding is defined as p_i = (i/s) * w_i with w_i = Norm(1 + alpha * |\nabla H_i|), which requires the gradient of the H&E image of the missing i-th slice. However, the inputs listed in Section 2 are only the adjacent ST slices I_0, I_1 and their paired H&E images H_0, H_1; no procedure is given for obtaining \nabla H_i from H_0 and H_1 alone. If the intermediate H&E image is available at inference time, then the claimed setting of interpolating missing slices from adjacent slices is violated and target information leaks into p_i; if it is not available, the module is undefined. This issue blocks evaluation of the central claim of arbitrary-position interpolation.
  2. [Section 2.3, position set P] The text states that the model generates s imputed intermediate ST slices, but the position set is defined as P = {p_i | p_i = (i/s) * w_i, i values from 1 to s-1}, which contains only s-1 positions. Moreover, with i ranging from 1 to s-1, the values i/s produce positions 1/s, ..., (s-1)/s, which does not correspond to s intermediate locations between the endpoints 0 and 1; a spacing of i/(s+1) would be the natural choice. This internal inconsistency makes it unclear how the s output slices are generated and how the loss terms in Section 2.4 are computed.
  3. [Section 3.1, Table 2] All multi-slice experiments use s = 2, 3, or 4 evenly spaced intermediate slices, but no experiment varies the relative positions of the missing slices. The claim that C2-STi supports interpolation at arbitrary positions is therefore not empirically demonstrated. In addition, the paper reports results on a single split of a single dataset without error bars or significance tests, and several margins over baselines are small (e.g., SSIM 0.78 vs. 0.77 in the 2-slice row of Table 2), so the robustness of the headline empirical claim is unclear.
  4. [Section 2.2, Eq. (3)] The GCN update C_{L,graph}^0 = delta(P C_L^0 W_n) uses a propagation operator P that is never defined. It should be stated explicitly (for example, as a normalized adjacency matrix derived from the correlation matrix A in Eq. (2)) for the method to be reproducible and for the role of A to be understood.
minor comments (5)
  1. [Section 2.3, first paragraph] There is a duplicated word in 'to to quantify biological structure similarity', and the sentence defining d_{i,j} = (s+1)Delta_d is confusing because the right-hand side does not depend on i and j; the distance between slices I_i and I_j should presumably be (j-i)Delta_d.
  2. [Section 3, Dataset] The dataset description says '19 consecutive slices spanning 350 mm'; this is likely a unit error and should be micrometers, since 350 mm would be anatomically implausible for HNSCC tissue sections.
  3. [Section 2.2, MGC-Graph] The description of the MGC-Graph construction is ambiguous: it says the output features C_L^0 and C_L^1 are used as input nodes, but it is not clear how individual genes are identified as nodes in a graph of size N, nor how the same adjacency A is applied to both the forward and backward branches.
  4. [Section 3.1, Table 1] The experimental protocol for the baselines is underdescribed; please state whether the same training/validation split, the same number of input gene channels, and the same optimizer settings were used for RIFE, RBI, DAIN, IFRNet, and U-Net, and clarify how image-quality metrics are computed over genes.
  5. [References] Several references cite only arXiv versions without DOIs; adding published versions where available would improve the reference list.

Circularity Check

1 steps flagged · score 6.0 of 10

DLSM position encoding is defined from the H&E image of the very slice to be predicted, so the arbitrary-position mechanism is not derivable from the two adjacent slices alone.

  1. self definitional [Section 2.3, 'Cross-section distance modeling'; inputs specified in Section 2.1]
    "C2-STi takes as input two adjacent ST slices ... I0 and I1, along with their corresponding H&E images H0 and H1. ... the Sobel operator [4] is used to calculate the gradient-based weight of i-th slice as w_i = Norm(1+α|∇H_i|), where α is a gradient scaling parameter. Thus, the position set of the under-interpolated slices is obtained as P={p_i | p_i = i/s·w_i∈(0,1), i values from 1 to s−1}."

    The stated model inputs are only the two adjacent ST-H&E pairs, but p_i is computed from ∇H_i, the H&E of the target slice itself. Thus the target slice's own image defines the position at which its ST map is generated. For a genuinely missing section H_i does not exist, so the equation cannot be evaluated from the declared inputs; if H_i is loaded from the dataset during training/testing, the 'missing ST slice' is partially observed via its paired H&E, so the arbitrary-position derivation reduces to conditioning on target-slice information rather than using only adjacent slices.

full rationale

The benchmark comparison is otherwise self-contained: C2-STi is trained on training patches and tested on held-out test patches against external baselines (IFRNet, RBI, DAIN, RIFE, U-Net), so the quantitative claims are not circular. The self-citations (refs. 6, 19, 23) are background or architectural precedents and are not used to establish the main superiority claim; none is an imported uniqueness theorem or an ansatz that carries the derivation. The central circularity is the DLSM position set P: p_i = (i/s)·Norm(1+α|∇H_i|) requires the H&E gradient of the under-interpolated (target) slice, while the model's declared inputs are only H0 and H1 from adjacent slices. The paper neither explains how H_i is obtained at inference for genuinely missing sections nor restricts the experiments to a setting where H_i is hidden. This is a target-dependent conditioning variable inside the mechanism that is supposed to make arbitrary-position interpolation possible, so the main methodological derivation is partially self-referential. The internal index inconsistency (P contains s−1 positions while the task states s imputed slices, and i/s does not place s interior slices between endpoints) reinforces that this part of the derivation is not well-defined as written.

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

The central claim depends on the dataset's aligned pairs, on an untested mapping from H&E gradients to inter-slice position, and on several unreported hyperparameters. The only explicitly invented things are software modules; the paper introduces no new biological or physical entities.

free parameters (4)
  • lambda (graph balancing weight)
    Chosen by hand to blend graph-convolved and original node features in Eq. (3); no value is reported in the paper.
  • alpha (gradient scaling weight)
    Scales the H&E gradient contribution to the slice-position weight w_i in Section 2.3; no value is reported.
  • beta and gamma (channel/spatial fusion weights)
    Combine channel-adaptive and spatial-adaptive modulated features as F_out = beta*F_ca + gamma*F_sp; values are not reported.
  • lambda_sim and lambda_smo (loss weights) = 1 and 1
    Set to 1 in Section 2.4; chosen by hand rather than tuned. Minor because both losses are fixed.
assumptions (4)
  • domain assumption Adjacent ST slices are spatially aligned and correspond one-to-one at the patch level.
    The whole input/output construction in Section 3 assumes the 1st and 3rd slices, plus H&E pairs, are aligned so pixel-wise interpolation is meaningful.
  • domain assumption H&E-derived gradient magnitude is a valid proxy for inter-slice tissue deformation and for choosing interpolation positions.
    Section 2.3 defines w_i = Norm(1 + alpha*|gradient H_i|) and uses it to set p_i; no evidence in the paper validates this mapping.
  • domain assumption Optical-flow-style deformable fields can represent the biological change between tissue sections.
    The deformation module adopts video frame interpolation machinery (deformable convolution, offsets, masks) without checking whether tissue deformation is well modeled as dense pixel displacement.
  • ad hoc to paper The undefined propagation operator P in the GCN update is assumed to be a normalized adjacency derived from A, though this is never stated.
    Equation (3) uses P C_L^0 W_n after defining the co-expression matrix A; P is never defined, so the reader must supply an assumption to make the equation computable.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Adaptive Spatial Transcriptomics Interpolation via Cross-modal Cross-slice Modeling." pith.science (2026). https://pith.science/paper/I627W7LV

@misc{pith2026250510729,
  author       = {Pith},
  title        = {Pith review of: Adaptive Spatial Transcriptomics Interpolation via Cross-modal Cross-slice Modeling},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/I627W7LV}},
  note         = {Machine review of arXiv:2505.10729}
}
read the original abstract

Spatial transcriptomics (ST) is a promising technique that characterizes the spatial gene profiling patterns within the tissue context. Comprehensive ST analysis depends on consecutive slices for 3D spatial insights, whereas the missing intermediate tissue sections and high costs limit the practical feasibility of generating multi-slice ST. In this paper, we propose C2-STi, the first attempt for interpolating missing ST slices at arbitrary intermediate positions between adjacent ST slices. Despite intuitive, effective ST interpolation presents significant challenges, including 1) limited continuity across heterogeneous tissue sections, 2) complex intrinsic correlation across genes, and 3) intricate cellular structures and biological semantics within each tissue section. To mitigate these challenges, in C2-STi, we design 1) a distance-aware local structural modulation module to adaptively capture cross-slice deformations and enhance positional correlations between ST slices, 2) a pyramid gene co-expression correlation module to capture multi-scale biological associations among genes, and 3) a cross-modal alignment module that integrates the ST-paired hematoxylin and eosin (H&E)-stained images to filter and align the essential cellular features across ST and H\&E images. Extensive experiments on the public dataset demonstrate our superiority over state-of-the-art approaches on both single-slice and multi-slice ST interpolation. Codes are available at https://github.com/XiaofeiWang2018/C2-STi.

Figures

Figures reproduced from arXiv: 2505.10729 by the authors.

Figure 1
Figure 1. Framework of the C2-STi, including three modules of cross-modal alignment, pyramid gene co-expression correlation and distance-aware local structural modulation. Specifically, We design a pyramid gene co-expression correlation module that in￾tegrates a pyramid encoder, a multi-gene co-expression graph (MGC-Graph), and a multi-scale decoder, capturing deformation information between slices at various scales while ref… view at source ↗
Figure 2
Figure 2. Visual comparisons for single slice interpolation result. Note that IGKC, MT￾CO2, MT-CO3, and MT-ND2 denote different genes. Implementation details. Our method is optimized using AdamW [10] and im￾plemented with PyTorch. We trained our model for 40 epochs on seven NVIDIA RTX V100 32 GB GPUs, with a weight decay of 10−4 . The initial learning rate is set to 10−4 , which is decreased to 10−6 using cosine annealing. Th… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

24 extracted references · 16 canonical work pages

  1. [1]

    Bao, W., Lai, W.S., Ma, C., Zhang, X., Gao, Z., Yang, M.H.: Depth-aware video frame interpolation (2019),https://arxiv.org/abs/1904.00830

    work page Pith review arXiv 2019

  2. [2]

    IEEE Transactions on Medical Imaging27(3), 370–377 (2008)

    Frakes, D.H., Dasi, L.P., Pekkan, K., Kitajima, H.D., Sundareswaran, K., Yo- ganathan, A.P., Smith, M.J.T.: A new method for registration-based medical im- age interpolation. IEEE Transactions on Medical Imaging27(3), 370–377 (2008). https://doi.org/10.1109/TMI.2007.907324

    arXiv 2008

  3. [3]

    Huang, Z., Zhang, T., Heng, W., Shi, B., Zhou, S.: Real-time intermediate flow estimation for video frame interpolation (2022),https://arxiv.org/abs/2011. 06294

    work page 2022

  4. [4]

    IEEE Journal of solid-state circuits23(2), 358–367 (1988)

    Kanopoulos, N., Vasanthavada, N., Baker, R.L.: Design of an image edge detection filter using the sobel operator. IEEE Journal of solid-state circuits23(2), 358–367 (1988)

    work page 1988

  5. [5]

    Kong, L., Jiang, B., Luo, D., Chu, W., Huang, X., Tai, Y., Wang, C., Yang, J.: Ifrnet: Intermediate feature refine network for efficient frame interpolation (2022), https://arxiv.org/abs/2205.14620

    work page Pith review arXiv 2022

  6. [6]

    International Journal of Imaging Systems and Technology34(5), e23161 (2024)

    Li, M., Que, N., Zhang, J., Du, P., Dai, Y.: Vtcnet: A feature fusion dl model based on cnn and vit for the classification of cervical cells. International Journal of Imaging Systems and Technology34(5), e23161 (2024)

    work page 2024

  7. [7]

    Expert Systems with Applications272, 126602 (2025).https://doi.org/https://doi.org/10.1016/ j.eswa.2025.126602,https://www.sciencedirect.com/science/article/pii/ S0957417425002246

    Li, W., Song, H., Ai, D., Shi, J., Fan, J., Xiao, D., Fu, T., Lin, Y., Wu, W., Yang, J.: Sfcli-net: Spatial-frequency collaborative learning interpola- tion network for computed tomography slice synthesis. Expert Systems with Applications272, 126602 (2025).https://doi.org/https://doi.org/10.1016/ j.eswa.2025.126602,https://www.sciencedirect.com/science/ar...

    work page arXiv 2025

  8. [8]

    Nature Communications15(1), 600 (January 2024).https://doi.org/10.1038/s41467-024-44835-w,https://doi

    Liang, Y., Shi, G., Cai, R., Yuan, Y., Xie, Z., Yu, L., Huang, Y., Shi, Q., Wang, L., Li, J., Tang, Z.: Prost: quantitative identification of spatially variable genes and domain detection in spatial transcriptomics. Nature Communications15(1), 600 (January 2024).https://doi.org/10.1038/s41467-024-44835-w,https://doi. org/10.1038/s41467-024-44835-w

    work page doi:10.1038/s41467-024-44835-w 2024

Show all 24 references
  1. [9]

    IEEE Transactions on Image Processing21(4), 1500–1512 (2011)

    Liu, A., Lin, W., Narwaria, M.: Image quality assessment based on gradient simi- larity. IEEE Transactions on Image Processing21(4), 1500–1512 (2011)

    2011

  2. [10]

    arXiv preprint arXiv:1711.05101 (2017)

    Loshchilov, I., Hutter, F.: Decoupled weight decay regularization. arXiv preprint arXiv:1711.05101 (2017)

    2017 arXiv

  3. [11]

    Scientific Reports12(1), 4133 (2022)

    Monjo, T., Koido, M., Nagasawa, S., Suzuki, Y., Kamatani, Y.: Efficient prediction of a spatial transcriptomics profile better characterizes breast cancer tissue sections without costly experimentation. Scientific Reports12(1), 4133 (2022)

    2022

  4. [12]

    Nature methods19(5), 534–546 (2022)

    Moses, L., Pachter, L.: Museum of spatial transcriptomics. Nature methods19(5), 534–546 (2022)

    2022

  5. [13]

    Park, S., Han, D., Kwak, N.: The u-net based glow for optical-flow-free video interframe generation (2021),https://arxiv.org/abs/2103.09576

    2021 arXiv

  6. [14]

    IEEE Transactions on Medical Imaging23(7), 922–926 (2004)

    Penney, G., Schnabel, J., Rueckert, D., Viergever, M., Niessen, W.: Registration- based interpolation. IEEE Transactions on Medical Imaging23(7), 922–926 (2004). https://doi.org/10.1109/TMI.2004.828352

    2004

  7. [15]

    Cell187(26), 7351–7373 (2024)

    Qiu, X., Zhu, D.Y., Lu, Y., Yao, J., Jing, Z., Min, K.H., Cheng, M., Pan, H., Zuo, L., King, S., et al.: Spatiotemporal modeling of molecular holograms. Cell187(26), 7351–7373 (2024)

    2024

  8. [16]

    Cell 187(15), 3953–3972.e26 (July 2024).https://doi.org/10.1016/j.cell.2024

    Schott, M., León-Periñán, D., Splendiani, E., Strenger, L., Licha, J.R., Penti- malli, T.M., Schallenberg, S., Alles, J., Tagliaferro, S.S., Boltengagen, A., Ehrig, Arbitrary Slice Spatial Transcriptomics Interpolation 11 S., Abbiati, S., Dommerich, S., Pagani, M., Ferretti, E...

    2024 doi

  9. [17]

    Shi, Z., Liu, X., Shi, K., Dai, L., Chen, J.: Video frame interpolation via generalized deformable convolution (2021),https://arxiv.org/abs/2008.10680

    2021 arXiv

  10. [18]

    Nature Biotechnology41(6), 773–782 (2023)

    Tian, L., Chen, F., Macosko, E.Z.: The expanding vistas of spatial transcriptomics. Nature Biotechnology41(6), 773–782 (2023)

    2023

  11. [19]

    In: International Conference on Medical Image Computing and Computer-Assisted Intervention

    Wang, X., Huang, X., Price, S., Li, C.: Cross-modal diffusion modelling for super- resolved spatial transcriptomics. In: International Conference on Medical Image Computing and Computer-Assisted Intervention. pp. 98–108. Springer (2024)

    2024

  12. [20]

    arXiv preprint arXiv:2303.14845 (2023)

    Wang, X., Price, S., Li, C.: Multi-task learning of histology and molecular markers for classifying diffuse glioma. arXiv preprint arXiv:2303.14845 (2023)

    2023 arXiv

  13. [21]

    In: Proceedings of the 47th International Conference on Parallel Processing

    You, Y., Zhang, Z., Hsieh, C.J., Demmel, J., Keutzer, K.: Imagenet training in minutes. In: Proceedings of the 47th International Conference on Parallel Processing. ICPP ’18, Association for Computing Machinery, New York, NY, USA(2018).https://doi.org/10.1145/3225058.3225069,h...

    2018

  14. [22]

    Nature Biotechnology42(9), 1372–1377 (Sep 2024).https://doi.org/ 10.1038/s41587-023-02019-9,https://doi.org/10.1038/s41587-023-02019-9

    Zhang, D., Schroeder, A., Yan, H., Yang, H., Hu, J., Lee, M.Y.Y., Cho, K.S., Susz- tak, K., Xu, G.X., Feldman, M.D., Lee, E.B., Furth, E.E., Wang, L., Li, M.: Infer- ring super-resolution tissue architecture by integrating spatial transcriptomics with histology. Nature Biotech...

    2024 doi

  15. [23]

    In: International Conference on Medical Image Computing and Computer-Assisted Intervention

    Zhang, Y., Wang, X., Meng, F., Tang, J., Li, C.: Knowledge-driven subspace fusion and gradient coordination for multi-modal learning. In: International Conference on Medical Image Computing and Computer-Assisted Intervention. pp. 263–273. Springer (2024)

    2024

  16. [24]

    Zhao, Q., Zhang, Q.: High-dimensional bayesian model for disease-specific gene detection in spatial transcriptomics (2024),https://arxiv.org/abs/2409.02397

    2024 arXiv

Pith tools

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