CoChat

Non-linear Characterization of Commercial and Decellularized Hydrogels: Statistical Framework Enhanced by Bayesian Optimization

Open access
García-García, D. E., Marques, D., Amaveda, H., Mora, M., Asín, J., Villaoslada, I., Baptista, P. M., Pérez, M. A., García-Aznar, J. M. · 2026 · Cellular and Molecular Bioengineering

The DOI matches its Crossref record; 63 of the 63 references it cites were found in a registry or an index.

Abstract

Scope: Hydrogels are widely used in the design of tissue substitutes because of their ability to mimic the extracellular matrix (ECM). Their mechanical cues critically influence the cellular response, making accurate characterization essential. However, it remains challenging due to their intricate nature. This study computationally evaluates the hyperelastic properties of next-generation hydrogels of high biomedical interest, including basal membrane extract and decellularized liver matrices, as well as structural proteins. Methods: We present a combined framework based on Bayesian optimization and statistical analyses that go beyond classical least-squares fitting, leveraging rheological experimental data. It defines each hyperelastic strain-energy density function, and addresses both intra- and inter-sample variability. This approach quantifies uncertainty and reveals the natural variability that deterministic models overlook, and it also enables quantification of coefficient variation with composition. Results: Validation against experimental data shows computational fits of 5% error in most cases, and low calculation time. Analyses reveal that composition—collagen addition, fibrin concentration, and decellularized extracellular matrix (dECM) age—modulate initial stiffness, nonlinearity, and overall mechanical resistance of the hydrogels. Conclusion: Hydrogels derived from basal membrane extracts exhibit comparable non-linear mechanics, while collagen addition reduces stiffness and nonlinearity. In fibrin-based hydrogels with decellularized liver matrix, mechanical behavior is concentration and age-dependent. Such insights are highly relevant for mechanobiology, enabling prediction of how cells sense mechanical cues and scaffold composition influences in vivo interactions.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1007/s12195-026-00913-1

Cites

63 of 63 references verified0 citation errors · 0 not checkable

Statistics

Figures

Fig. 1 Workflow diagram illustrating the main steps performed. Sample preparation includes basal membrane extract-based hydrogel formulations (BME, Matrigel and BME combined with collagen type I), as well as fibrin-based hydrogels at 3 mg/mL and 5.75 mg/mL concentrations. The last ones are also combined with decellularized liver matrices, fetal or adult, obtained through a decellularization process. Each material is characterized by rheology tests. Rheological data are preprocessed to isolate the region of interest, corresponding to elastic behavior. A preliminary evaluation identifies the most suitable ψ to describe the mechanical behavior of the hydrogels, the second-order (N = 2) Polynomial function. Next, the simplified FEM proposed by Valero et al. [25] is implemented to compute the stress–strain curves, applying a displacement representing half of the loading cycle. To identify the coefficients of the N = 2 Polynomial function that describes each experimental behavior, a Gaussian-process-based Bayesian optimization is employed. Finally, two statistical analyses to address intra- and inter-material variability of the hydrogels are conducted
Fig. 1 Workflow diagram illustrating the main steps performed. Sample preparation includes basal membrane extract-based hydrogel formulations (BME, Matrigel and BME combined with collagen type I), as well as fibrin-based hydrogels at 3 mg/mL and 5.75 mg/mL concentrations. The last ones are also combined with decellularized liver matrices, fetal or adult, obtained through a decellularization process. Each material is characterized by rheology tests. Rheological data are preprocessed to isolate the region of interest, corresponding to elastic behavior. A preliminary evaluation identifies the most suitable ψ to describe the mechanical behavior of the hydrogels, the second-order (N = 2) Polynomial function. Next, the simplified FEM proposed by Valero et al. [25] is implemented to compute the stress–strain curves, applying a displacement representing half of the loading cycle. To identify the coefficients of the N = 2 Polynomial function that describes each experimental behavior, a Gaussian-process-based Bayesian optimization is employed. Finally, two statistical analyses to address intra- and inter-material variability of the hydrogels are conductedCC BY · PubMed Central
Fig. 2 Bayesian optimization algorithm flowchart. The iterative workflow includes coefficient selection within the predefined search space, computation of the computational curve and error metric (MSE), and guidance of subsequent iterations by the Bayesian optimization algorithm
Fig. 2 Bayesian optimization algorithm flowchart. The iterative workflow includes coefficient selection within the predefined search space, computation of the computational curve and error metric (MSE), and guidance of subsequent iterations by the Bayesian optimization algorithmCC BY · PubMed Central
Fig. 3 Comparison between experimental and simulated τ (MPa)–γ (–) curves for different samples of basal membrane extract-based hydrogels. For each pair of curves there are indicated the MSE values obtained using the optimization algorithm, indicating the quality of the fit. All plots are shown at the same scale; zoomed-in insets are provided in plots where the fitting details are unclear. There are shown the following materials: a BME, b Matrigel (with an offset of 2.00×10⁻⁵ for the first curve set), and c BME with Collagen Type I (with an offset of -2.00×10⁻⁶ units for the third). Offsets are applied for improved visualization. d and e Display the results of the statistical analysis conducted to generalize the mechanical behavior of BME, “Derivation of Generic Hyperelastic Function” section. Both plots show the mean τ–γ curve along with the variability range given by the individual material responses. d Corresponds to the regression-based approach, including MSE values. e Shows the results from the statistical normalization method. The shaded areas represent the 95% PIs in d, and the 80% FIs in e
Fig. 3 Comparison between experimental and simulated τ (MPa)–γ (–) curves for different samples of basal membrane extract-based hydrogels. For each pair of curves there are indicated the MSE values obtained using the optimization algorithm, indicating the quality of the fit. All plots are shown at the same scale; zoomed-in insets are provided in plots where the fitting details are unclear. There are shown the following materials: a BME, b Matrigel (with an offset of 2.00×10⁻⁵ for the first curve set), and c BME with Collagen Type I (with an offset of -2.00×10⁻⁶ units for the third). Offsets are applied for improved visualization. d and e Display the results of the statistical analysis conducted to generalize the mechanical behavior of BME, “Derivation of Generic Hyperelastic Function” section. Both plots show the mean τ–γ curve along with the variability range given by the individual material responses. d Corresponds to the regression-based approach, including MSE values. e Shows the results from the statistical normalization method. The shaded areas represent the 95% PIs in d, and the 80% FIs in eCC BY · PubMed Central
Fig. 4 Comparison between experimental and simulated τ (MPa)–γ (–) curves for different samples of fibrin-based hydrogels with or without decellularized liver matrix. For each pair of curves, the MSE values obtained from the optimization algorithm are indicated, reflecting the quality of the fit. All plots are shown at the same scale; zoomed-in insets are provided in plots where the fitting details are unclear. Subfigures a–c Correspond to hydrogels with a fibrin concentration of 3 mg/mL—shown in shades of green—without decellularized matrix, with ADM and FDM (offsets of 5.00×10⁻⁷ and 1.40×10⁻⁶ units for the second and third sets), respectively. Subfigures d–f show 5.75 mg/mL fibrin hydrogels with the same dECM compositions—shown in shades of blue. Offsets were also applied to ADM (5.00×10⁻⁶ and 1.00×10⁻⁵ for the second and third sets) and FDM (5.00×10⁻⁷ and 1.00×10⁻⁶, respectively). Offsets are applied for better visualization
Fig. 4 Comparison between experimental and simulated τ (MPa)–γ (–) curves for different samples of fibrin-based hydrogels with or without decellularized liver matrix. For each pair of curves, the MSE values obtained from the optimization algorithm are indicated, reflecting the quality of the fit. All plots are shown at the same scale; zoomed-in insets are provided in plots where the fitting details are unclear. Subfigures a–c Correspond to hydrogels with a fibrin concentration of 3 mg/mL—shown in shades of green—without decellularized matrix, with ADM and FDM (offsets of 5.00×10⁻⁷ and 1.40×10⁻⁶ units for the second and third sets), respectively. Subfigures d–f show 5.75 mg/mL fibrin hydrogels with the same dECM compositions—shown in shades of blue. Offsets were also applied to ADM (5.00×10⁻⁶ and 1.00×10⁻⁵ for the second and third sets) and FDM (5.00×10⁻⁷ and 1.00×10⁻⁶, respectively). Offsets are applied for better visualizationCC BY · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 63 references

Research works printed but not linked: 8

  1. Lanza, R., R. Langer, J. P. Vacanti, and A. Atala. Principles of Tissue Engineering. New York: Academic, 2020.
  2. Drew, D. A., and S. L. Passman. Theory of Multicomponent Fluids, Vol. 135, Berlin: Springer, 2006.
  3. Abaqus Scripting User’s Manual. Abaqus 6.11. http://130.149.890(2080):0v.6, 2012.
  4. Martínez, F. M.. Determining the Biomechanical Behavior of the Liver Using Medical Image Analysis and Evolutionary Computation. PhD Thesis, Universitat Politècnica de València, 2014.
  5. Snoek, J., H. Larochelle, and R. P. Adams. Practical Bayesian optimization of machine learning algorithms. In: Advances in Neural Information Processing Systems, 2012, p. 25.
  6. Montgomery, D. C., E. A. Peck, and G. G. Vining. Introduction to Linear Regression Analysis. Hoboken: Wiley, 2021.
  7. Abdi, H., and L. J. Williams. Newman–Keuls test and Tukey test. Encyclop. Res. Des. 2:897–902, 2010.
  8. Tapia-Romero, M. A., M. Dehonor-Gómez, and L. E. Lugo-Uribe. Prony series calculation for viscoelastic behavior modeling of structural adhesives from DMA data. Ingen. invest. tecnol. 21(2):1–10, 2020.

Cited by

Still being checked.

Metadata

Authors
García-García, D. E., Marques, D., Amaveda, H., Mora, M., Asín, J., Villaoslada, I., Baptista, P. M., Pérez, M. A., & García-Aznar, J. M.
Affiliations
  • Universidad de Zaragoza (ES)
  • Fundacion Agencia Aragonesa para la Investigacion y el Desarrollo (ES)
Venue
Cellular and Molecular Bioengineering
Published
2026-06
Volume
19
Issue
3
Pages
327-344
ISSN
  • 1865-5025 (ISSN-L)
  • 1865-5033
OpenAlex
W7166655945
Cited by
0
References
61
FWCI
0.00
Type
journal article
Language
en
Open access
Licence
Locations
  • Cellular and Molecular Bioengineering (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • PubMed Central (publishedVersion, cc-by)
  • Zaguan (University of Zaragoza Repository) (submittedVersion, cc-by)
  • DIGITAL.CSIC (Spanish National Research Council (CSIC)) (submittedVersion, other-oa)
Topics
  • Physical Sciences › Materials Science › Biomaterials › Hydrogels: synthesis, properties, applications
  • Physical Sciences › Engineering › Biomedical Engineering › Elasticity and Material Modeling
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Cell Biology › Cellular Mechanics and Interactions
Keywords
hydrogels, hyperelasticity, strain energy function, Bayesian optimization, decellularized extracellular matrix, fibrin gel, hydrogel stiffness, nonlinearity, uncertainty quantification, tissue substitutes, mechanobiology

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Non-linear Characterization of Commercial and Decellularized Hydrogels: Statistical Framework Enhanced by Bayesian Optimization. https://app.cochat.ai/papers/10.1007/s12195-026-00913-1
The paper
García-García, D. E., Marques, D., Amaveda, H., Mora, M., Asín, J., Villaoslada, I., Baptista, P. M., Pérez, M. A., & García-Aznar, J. M. (2026). Non-linear Characterization of Commercial and Decellularized Hydrogels: Statistical Framework Enhanced by Bayesian Optimization. Cellular and Molecular Bioengineering. https://doi.org/10.1007/s12195-026-00913-1

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.