OSCR

Self-organization of vascularized muscle from bovine embryonic stem cells.

Code ↔ Paper

4 matches between paragraphs of the paper and lines of its authors' code, computed by the harvester (lexical-v1). Click a colored paragraph or line to see its counterpart.

The 4 matches
  1. [1] § Methods › Endothelial network connectivity analysis ↔ src/model_utils.py, lines 109–155 · score 0.74 · RAdam, dice, IoU, loss, scheduler, model
  2. [2] § Methods › α-BTX quantification ↔ nmj_image_analysis.py, lines 239–300 · score 0.71 · distance transforms, neuron mask, muscle mask, EDTs, raw, pixel
  3. [3] § Methods › Endothelial network connectivity analysis ↔ src/training.py, lines 5–35 · score 0.59 · trained, IoU, loss, scheduler, encoder, model
  4. [4] § Methods › α-BTX quantification ↔ BTX_batch.py, lines 255–307 · score 0.55 · background, EDTs, BTX signals, batch, conservative, Detection

Paper

Loaded from Europe PMC by your browser, not stored by OSCR: doi.org · Europe PMC

The paper is loaded when this pane is shown.

The authors' code

Python · 155 lines · 5.5 KB · CC-BY-NC-4.0 · 1 match

  1. import math
  2. import segmentation_models_pytorch as smp
  3. import torch
  4. from catalyst import utils
  5. from catalyst.contrib.nn import DiceLoss, RAdam
  6. from catalyst.dl import (CriterionCallback, DiceCallback, IouCallback,
  7. MetricAggregationCallback, SupervisedRunner)
  9. def adapt_input_conv(in_chans, conv_weight):
  10. """
  11. This function adapts the input channels of a convolutional layer's weights based on the number of input channels
  12. provided. It handles cases where the input channels are 1 (grayscale), 3 (RGB), or other values.
  13. The function ensures that the weight tensor is in the correct format for the given number of input channels and
  14. adjusts the weights accordingly.
  16. Args:
  17. in_chans (int): The number of input channels.
  18. conv_weight (torch.Tensor): The convolutional layer's weights.
  20. Returns:
  21. torch.Tensor: The adapted convolutional layer's weights.
  22. """
  23. conv_type = conv_weight.dtype
  24. conv_weight = conv_weight.float() # Some weights are in torch.half, ensure it's float for sum on CPU
  25. O, I, J, K = conv_weight.shape
  26. if in_chans == 1:
  27. if I > 3:
  28. assert conv_weight.shape[1] % 3 == 0
  29. # For models with space2depth stems
  30. conv_weight = conv_weight.reshape(O, I // 3, 3, J, K)
  31. conv_weight = conv_weight.sum(dim=2, keepdim=False)
  32. else:
  33. conv_weight = conv_weight.sum(dim=1, keepdim=True)
  34. elif in_chans != 3:
  35. if I != 3:
  36. raise NotImplementedError('Weight format not supported by conversion.')
  37. else:
  38. # NOTE this strategy should be better than random init, but there could be other combinations of
  39. # the original RGB input layer weights that'd work better for specific cases.
  40. repeat = int(math.ceil(in_chans / 3))
  41. conv_weight = conv_weight.repeat(1, repeat, 1, 1)[:, :in_chans, :, :]
  42. conv_weight *= (3 / float(in_chans))
  43. conv_weight = conv_weight.to(conv_type)
  45. return conv_weight
  47. def adapt_input_model(model):
  48. """
  49. Adapts first layer to take a specified number of input channels.
  51. Args:
  52. model: The segmentation model to be adapted.
  54. Returns:
  55. The adapted segmentation model.
  56. """
  57. # Adapt first layer to take 1 channel as input - timm approach = sum weights
  58. new_weights = adapt_input_conv(in_chans=1, conv_weight=model.encoder.patch_embed1.proj.weight)
  59. model.encoder.patch_embed1.proj = torch.nn.Conv2d(in_channels=1, out_channels=64, kernel_size=(7, 7), stride=(4, 4), padding=(3, 3))
  61. with torch.no_grad():
  62. model.encoder.patch_embed1.proj.weight = torch.nn.parameter.Parameter(new_weights)
  64. return model
  66. def build_model(model_str, encoder_str, in_channels=1, encoder_weights=None):
  67. """
  68. Provide a segmentation model with specified architecture and encoder.
  70. Args:
  71. model_str (str): Name of the segmentation model architecture.
  72. encoder_str (str): Name of the encoder used in the model.
  73. encoder_weights (str): Pretrained weights for the specified encoder.
  74. in_channels (int): Number of input channels.
  76. Returns:
  77. A PyTorch model instance with specified architecture and encoder.
  78. """
  80. # Create a dictionary mapping from string to actual SMP function
  81. ARCHITECTURES = {
  82. 'Unet': smp.Unet,
  83. 'Unet++': smp.UnetPlusPlus,
  84. 'MAnet': smp.MAnet,
  85. 'Linknet': smp.Linknet,
  86. 'FPN': smp.FPN,
  87. 'PSPNet': smp.PSPNet,
  88. 'PAN': smp.PAN,
  89. 'DeepLabV3': smp.DeepLabV3,
  90. 'DeepLabV3+': smp.DeepLabV3Plus
  91. }
  93. try:
  94. # Get constructor method for desired architecture
  95. constructor = ARCHITECTURES[model_str]
  97. if encoder_str.startswith('mit') and in_channels == 1:
  98. model = constructor(encoder_name=encoder_str, classes=1, in_channels=3, encoder_weights=encoder_weights)
  99. model = adapt_input_model(model)
  100. else:
  101. model = constructor(encoder_name=encoder_str, classes=1, in_channels=in_channels, encoder_weights=encoder_weights)
  103. return model
  105. except KeyError:
  106. raise ValueError(f'Model type "{model_str}" not understood. Valid options are: {list(ARCHITECTURES.keys())}')
  109. def compile_runner(learning_rate, model_str, encoder_str, encoder_weights, loss_weights, n_epochs, in_channels):
  111. criterion = {
  112. "dice": DiceLoss(),
  113. "bce": torch.nn.BCEWithLogitsLoss()
  114. }
  116. device = utils.get_device()
  117. print(f"Using device: {device}")
  119. callbacks = [
  121. CriterionCallback(
  122. input_key="mask",
  123. prefix="loss_dice",
  124. criterion_key="dice"
  125. ),
  127. CriterionCallback(
  128. input_key="mask",
  129. prefix="loss_bce",
  130. criterion_key="bce"
  131. ),
  133. MetricAggregationCallback(
  134. prefix="loss",
  135. mode="weighted_sum",
  136. metrics={
  137. "loss_dice": loss_weights[0],
  138. "loss_bce": loss_weights[1]
  139. },
  140. ),
  142. # metrics
  143. DiceCallback(input_key="mask"),
  144. IouCallback(input_key="mask", threshold=0.5)
  146. ]
  148. model = build_model(model_str, encoder_str, in_channels, encoder_weights)
  149. model_params = utils.process_model_params(model)
  151. optimizer = RAdam(model_params, lr=learning_rate, weight_decay=1e-4)
  152. scheduler = torch.optim.lr_scheduler.StepLR(optimizer, step_size=int(n_epochs/2), gamma=0.5, last_epoch=- 1, verbose=False)
  153. runner = SupervisedRunner(device=device, input_key="image", input_target_key="mask")
  155. return model, runner, criterion, optimizer, scheduler, callbacks

model_utils.py at commit daee164, under CC-BY-NC-4.0 · at the source

Overview

Authors: Marina Sanaki-Matsumiya1,2, Yuya Sanaki2, Casandra Villava1, Luca Rappez1, Nicola Gritti1, Fumio Nakaki1, James Sharpe1, Kristina Haase1, Jun Wu3,4,5, Miki Ebisuya1,6
ORCID iDs: Marina Sanaki-Matsumiya, Yuya Sanaki, Nicola Gritti, Fumio Nakaki, James Sharpe, Kristina Haase, Jun Wu, Miki Ebisuya
  1. European Molecular Biology Laboratory, EMBL Barcelona, C/ Dr. Aiguader, Barcelona, Spain
  2. Life Science Center for Survival Dynamics, Tsukuba Advanced Research Alliance (TARA), University of Tsukuba, Ibaraki, Japan
  3. Department of Molecular Biology, University of Texas Southwestern Medical Center, Dallas, TX USA
  4. Hamon Center for Regenerative Science and Medicine, University of Texas Southwestern Medical Center, Dallas, TX USA
  5. Cecil H. and Ida Green Center for Reproductive Biology Sciences, University of Texas Southwestern Medical Center, Dallas, TX USA
  6. Cluster of Excellence Physics of Life, TU Dresden, Dresden, Germany
Institutions: University of Tsukuba (Japan); European Molecular Biology Laboratory (Spain); Life Science Center for Survival Dynamics, Tsukuba Advanced Research Alliance (TARA); The University of Texas Southwestern Medical Center (United States); Physics of Life (Germany); Technische Universität Dresden (Germany)
Journal: Nature communications, volume 17, issue 1, article 8975
Dates: received 15 February 2024; accepted 30 July 2026; published online 2 September 2026
Type: Research article · Language: English
License: CC BY
Identifiers: DOI 10.1038/s41467-026-76569-2 · PMID 42686746 · PMCID PMC13538390 · OpenAlex W7204998249
Open access: gold, a free copy (OpenAlex)
Status: code verified
Categories: other (organism)
Methods: Smoothing, state filtering, decompositions, Preprocessing, Machine learning
Keywords: Stem-cell differentiation, Tissue engineering, Musculoskeletal development
MeSH: Embryonic Stem Cells*, Muscle, Skeletal*, Animals, Cattle, Cell Culture Techniques, Cell Differentiation, Cells, Cultured, Endothelial Cells, In Vitro Meat, Mesoderm, Muscle Fibers, Skeletal, Neurons, Red Meat (* major topic)
Topic: Pluripotent Stem Cells Research (Molecular Biology, Biochemistry, Genetics and Molecular Biology), according to OpenAlex
Funding: Welch Foundation (I-2088); Deutsche Forschungsgemeinschaft (German Research Foundation) (EXC 2068 - 390729961); European Research Council (101002564, 101040977); NIGMS NIH HHS (R01 GM138565); NICHD NIH HHS (R01 HD103627); U.S. Department of Health & Human Services | National Institutes of Health (NIH) (GM138565-01A1 and HD103627-01A1)
Citations: cited by 1 paper (Europe PMC); 69 references in the paper
Research resources: RRID:SCR_023307

Abstract

Cultured beef offers a promising alternative to traditional meat. While adult stem cells are commonly used as the cell source for cultured beef, their proliferation and differentiation capacities are limited. Current manufacturing processes for cultured beef steaks often require separate preparation of multiple cell types and their intricate assembly. In this study, we propose and report the co-induction of muscle, neural, and endothelial cells from bovine embryonic stem cells (ESCs), accompanied by the formation of tissue-like structures in two- and three-dimensional cultures. Bovine myocytes are induced in a stepwise manner through the induction of presomitic mesoderm (PSM) from bovine ESCs. Muscle fibers with sarcomeres appear within 15 days, displaying calcium oscillations responsive to inputs from co-induced bovine neurons. Bovine endothelial cells are also co-induced via PSM, forming uniform vessel networks inside tissues. Our serum-free, rapid co-induction protocols represent a milestone toward self-organizing beef steaks with integrated vasculature and innervation.

Reproduced under the paper's license (CC BY), from the paper cited above.

Repositories

Its files are read in the Code ↔ Paper reader above, with 4 matches between paragraphs and lines of code.

YuyaSanaki/NMJ-analysis

License: Apache-2.0
State: the link answers, verified on 26 September 2026
Evidence: files inventoried
Commit: 4c8fd8aa09233c648ea0ed8eb1586da001c2a8c3, 8 July 2026
Languages: Python (8), Shell (1)
Size: 24 files, 9 scripts
Software Heritage: not archived
Found in: “Code availability”
Holds: README, license file, environment (docker-compose.yml, Dockerfile, requirements.txt)
Not found: CITATION.cff, tests, continuous integration, documentation
Tools: Matplotlib (7 files), pandas (5 files), NumPy (4 files), scikit-image (3 files), SciPy (2 files), seaborn (2 files), tifffile (1 file)
Availability: 1 check, the latest on 26 September 2026: the link answers
  • 26 September 2026: the link answers
11 files
At the source: github.com/YuyaSanaki/NMJ-analysis

git.embl.de/grp-mif/image-analysis

License: none: the authors keep all their rights
State: the link answers, verified on 26 September 2026
Evidence: the link answers
Software Heritage: not checked
Found in: “Code availability”
Not found: README, license file, CITATION.cff, environment file, tests, continuous integration, documentation
Availability: 1 check, the latest on 26 September 2026: the link answers (HTTP 200)
  • 26 September 2026: the link answers (HTTP 200)
At the source: git.embl.de/grp-mif/image-analysis/meatball_analysis

LRpz/VascuMap

License: CC-BY-NC-4.0
State: the link answers, verified on 26 September 2026
Evidence: files inventoried
Commit: daee1647eaf4c78262bc68afd7e6f2d29beecf5c, 13 January 2025
Languages: Python (7)
Size: 16 files, 7 scripts
Software Heritage: not archived
Found in: “Code availability”
Holds: README, license file, environment (setup.py)
Not found: CITATION.cff, tests, continuous integration, documentation
Tools: NumPy (5 files), PyTorch (4 files), scikit-image (4 files), tifffile (4 files), Matplotlib (2 files), NetworkX (2 files), pandas (2 files), SciPy (2 files), scikit-learn (1 file)
Availability: 1 check, the latest on 26 September 2026: the link answers
  • 26 September 2026: the link answers
9 files
At the source: github.com/LRpz/VascuMap

Code availability

The custom scripts used are available from Github [https://git.embl.de/grp-mif/image-analysis/meatball_analysis], [https://github.com/LRpz/VascuMap], and [https://github.com/YuyaSanaki/NMJ-analysis].

Reproduced under the paper's license (CC BY), from the paper cited above.

Tracing map

Proposed by the machine: these links were found in the paper and verified at the source, without human review. The map will receive a Zenodo DOI once one of the paper's authors has validated it with their ORCID.

What the map holds:

  • 3 repositories of the authors' code, each at its verified commit, with its license and how the link was found in the paper;
  • 16 scripts, each with its path and the digest of its content;
  • 4 matches between paragraphs of the paper and lines of the code (method lexical-v1);
  • neither the text of the paper nor the code itself.

Its JSON (tracing-map.json) is deposited on Zenodo with its DOI once the map is validated.

Data

Datasets cited

Data Availability Statement

Source data are provided with this paper. The scRNA-seq data generated in this study have been deposited in the ArrayExpress database under accession code #E-MTAB-14714 (https://www.ebi.ac.uk/biostudies/arrayexpress/studies/E-MTAB-14714). Source data are provided with this paper.

The custom scripts used are available from Github [https://git.embl.de/grp-mif/image-analysis/meatball_analysis], [https://github.com/LRpz/VascuMap], and [https://github.com/YuyaSanaki/NMJ-analysis].

Reproduced under the paper's license (CC BY), from the paper cited above.

Versions

The history of this record: each version stored by the harvester or made by a correction of its authors or of the maintainers of its code, and what changed in its facts. The texts of the paper (its abstract, its availability statements) are not part of it; versions that changed only those are not listed.

Version 1, 27 September 2026: the first record

Recorded: type, language, journal, volume, issue, pages, dates, 10 authors, 3 keywords, 13 MeSH terms, 6 funders, 64 references, 1 RRID.

Cite

This paper

Sanaki-Matsumiya, M., Sanaki, Y., Villava, C., Rappez, L., Gritti, N., Nakaki, F., Sharpe, J., Haase, K., Wu, J., & Ebisuya, M. (2026). Self-organization of vascularized muscle from bovine embryonic stem cells. Nature communications, 17(1), 8975. https://doi.org/10.1038/s41467-026-76569-2

BibTeX

@article{sanakimatsumiya2026self,
author = {Sanaki-Matsumiya, Marina and Sanaki, Yuya and Villava, Casandra and Rappez, Luca and Gritti, Nicola and Nakaki, Fumio and Sharpe, James and Haase, Kristina and Wu, Jun and Ebisuya, Miki},
title = {{Self-organization of vascularized muscle from bovine embryonic stem cells}},
journal = {Nature communications},
year = {2026},
month = sep,
volume = {17},
number = {1},
pages = {8975},
publisher = {Nature Publishing Group},
issn = {2041-1723},
doi = {10.1038/s41467-026-76569-2},
url = {https://doi.org/10.1038/s41467-026-76569-2},
pmid = {42686746},
pmcid = {PMC13538390}
}

RIS

TY - JOUR
AU - Sanaki-Matsumiya, Marina
AU - Sanaki, Yuya
AU - Villava, Casandra
AU - Rappez, Luca
AU - Gritti, Nicola
AU - Nakaki, Fumio
AU - Sharpe, James
AU - Haase, Kristina
AU - Wu, Jun
AU - Ebisuya, Miki
TI - Self-organization of vascularized muscle from bovine embryonic stem cells
T2 - Nature communications
J2 - Nat Commun
PY - 2026
DA - 2026/09/02
VL - 17
IS - 1
SP - 8975
SN - 2041-1723
PB - Nature Publishing Group
DO - 10.1038/s41467-026-76569-2
UR - https://doi.org/10.1038/s41467-026-76569-2
LA - en
ER -

CSL-JSON

{
"id": "10.1038/s41467-026-76569-2",
"type": "article-journal",
"title": "Self-organization of vascularized muscle from bovine embryonic stem cells",
"container-title": "Nature communications",
"author": [
{
"family": "Sanaki-Matsumiya",
"given": "Marina"
},
{
"family": "Sanaki",
"given": "Yuya"
},
{
"family": "Villava",
"given": "Casandra"
},
{
"family": "Rappez",
"given": "Luca"
},
{
"family": "Gritti",
"given": "Nicola"
},
{
"family": "Nakaki",
"given": "Fumio"
},
{
"family": "Sharpe",
"given": "James"
},
{
"family": "Haase",
"given": "Kristina"
},
{
"family": "Wu",
"given": "Jun"
},
{
"family": "Ebisuya",
"given": "Miki"
}
],
"container-title-short": "Nat Commun",
"volume": "17",
"issue": "1",
"page": "8975",
"DOI": "10.1038/s41467-026-76569-2",
"PMID": "42686746",
"PMCID": "PMC13538390",
"ISSN": "2041-1723",
"publisher": "Nature Publishing Group",
"URL": "https://doi.org/10.1038/s41467-026-76569-2",
"language": "en",
"issued": {
"date-parts": [
[
2026,
9,
2
]
]
}
}

The tracing map gets a citation of its own once an author has validated it and it has a DOI.

Similar papers

The papers with a page that share the most with this one: the tools found in their code, their categories, datasets, cited references and authors, the rarest counting most.

[1] doi:10.64898/2026.03.30.715222 [code]
Synthetic lumen rounding directs neural progenitor division mode
Journal: bioRxiv (preprint)
In common: tifffile, scikit-image, PyTorch, 6 other tools, author Miki Ebisuya
[2] doi:10.1002/advs.202522762 [code]
Enhancing Maturation of Human Neuromuscular Organoids via Electrical Stimulation.
Journal: Advanced science (Weinheim, Baden-Wurttemberg, Germany)
In common: tifffile, scikit-image, seaborn, 4 other tools, 3 references
[3] doi:10.1371/journal.pcbi.1014571 [code]
SynAPSeg: A novel dataset and image analysis framework for deep learning-based synapse detection and quantification.
Journal: PLoS computational biology
In common: tifffile, NetworkX, scikit-image, 7 other tools
[4] doi:10.1016/j.isci.2026.116206 [code]
Gut distension evokes rapid neural dynamics in vagal and hindbrain populations of larval zebrafish.
Journal: iScience
In common: tifffile, NetworkX, scikit-image, 7 other tools
[5] doi:10.3389/fnsys.2026.1822122 [code]
Convergence-divergence circuits for multimodal integration of innate and learned opponent valences.
Journal: Frontiers in systems neuroscience
In common: tifffile, NetworkX, scikit-image, 7 other tools
[6] doi:10.1038/s41467-026-73373-w [code]
Mapping neuro-vascular unit communications reveals distinct angiogenic programs across developing mouse brain regions.
Journal: Nature communications
In common: tifffile, scikit-image, PyTorch, 5 other tools, 1 reference
[7] doi:10.1038/s41598-026-57519-w [code]
Automated segmentation of neurons and spinal cord structures in immunofluorescence images using SpineDL.
Journal: Scientific reports
In common: tifffile, NetworkX, scikit-image, 6 other tools
[8] doi:10.7554/elife.100880 [code]
An applicable and efficient retrograde monosynaptic circuit mapping tool for larval zebrafish.
Journal: eLife
In common: tifffile, NetworkX, scikit-image, 6 other tools
[9] doi:10.1016/j.xcrm.2026.102766 [code]
A longitudinal single-cell and spatial multiomic atlas of pediatric high-grade glioma.
Journal: Cell reports. Medicine
In common: tifffile, NetworkX, scikit-image, 6 other tools
[10] doi:10.1007/s12021-026-09817-x [code]
Circle of Willis-Guided Localization for Simultaneous Detection and Classification of Large Vessel Occlusions in Brain CTA.
Journal: Neuroinformatics
In common: tifffile, scikit-image, PyTorch, 6 other tools

Contribute

The authors of this paper can claim it, correct its record and validate its tracing map, and the maintainers of its code (its owner, or a public member of its organization) correct what it says of their repository; anyone signed in can ask for its removal. Every request goes to OSCR's own machine, which answers it; your account page follows them.

Sign in with ORCID to claim this paper as one of its authors, correct its record or validate its tracing map: when the paper's metadata lists your ORCID iD, you are recognized at once. Maintainers of its code: sign in with GitHub, then claim the repository on your account page.

Request its removal

To ask OSCR to remove this record, the copies of its authors' scripts or its tracing map, use the removal request page: signed in, you say who you are, what to remove and why, then review and confirm the request. Published rules decide every request (how).

Discussion, reproductions, activity

Discussion: questions and error reports about this paper and its code, from signed-in readers and its authors. It opens with sign-in.

Reproductions: reports from readers who ran the authors' code: what they reproduced, with which environment, commit and data. It opens with sign-in.

Activity: what happens around this paper: new versions of its record, its map's validation, discussions and reproductions. It opens with sign-in.