Pith. sign in

REVIEW 3 major objections 5 minor 36 references

Latent Space-Driven Quantification of Biofilm Formation using Time Resolved Droplet Microfluidics

T0 review · 3 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read The paper claims that a variational autoencoder trained on just two images per setup can segment and quantify biofilm, aggregate, and patch areas in bright-field droplet images, revealing the biofilm lifecycle and treatment effects…

desk verdict The droplet platform and time-resolved imaging are genuinely useful; the VAE segmentation that drives every quantitative claim is unvalidated and manually tuned, so the 'accurate quantification' headline does not hold as written. read the letter →

arxiv 2507.07632 v1 pith:ENJTY2MM submitted 2025-07-10 physics.bio-ph

Daniela Pérez Guerrero , Jesús Manuel Antúnez Domínguez , Aurélie Vigne , Daniel Midtvedt , Wylie Ahmed , Lisa D. Muiznieks , Giovanni Volpe , Caroline Beck Adiels This is my paper · ORCID
classification physics.bio-ph
keywords BiofilmformationUnsupervisedsegmentationDropletmicrofluidicsMicroscopyVariationalautoencoderHigh-throughputscreeningBacillussubtilisTime-lapseimaging
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

This paper integrates a droplet-microfluidics platform with an unsupervised variational-autoencoder (VAE) image-analysis tool to quantify biofilm formation in situ. The central claim is that a VAE trained on just two grayscale images per microscopy setup—one showing a structured biofilm and one showing dispersed planktonic bacteria—creates a latent space in which a single manually selected threshold separates bacterial structures from background, enabling automated measurement of biofilm, aggregate, and patch areas over time. Such a label-free readout matters because most biofilm assays are bulk, destructive, or manually annotated, whereas here each droplet is an independent microenvironment that can be followed continuously and compared side by side under different treatments. The authors apply the pipeline to antibiotic sensitivity screening with fluorescence and to glycerol-promoted biofilm development with bright-field time-lapse, reporting that sub-inhibitory antibiotic concentrations enhance biofilm formation and that glycerol accelerates and synchronizes biofilm growth. The result is a scalable, label-free route to studying biofilm dynamics, dispersal, and possible regrowth in confined volumes.

What carries the argument

The load-bearing object is the one-dimensional latent space of a variational autoencoder, bypassing the decoder entirely. Input crops are converted to a radial/angular representation that makes learning rotation-invariant, so the same local structure is encoded the same way regardless of orientation. A manually chosen threshold on the latent variable separates background from bacterial content, and applying that threshold on a structured grid over each droplet turns the latent map into segmentation masks and area measurements. The classifier is therefore not learned in the usual sense: the VAE learns a compressed visual code, but the background/bacteria decision is a human-set scalar cutoff, and the aggregate/biofilm/patch labels are imposed by pixel-area thresholds.

What would settle it

A concrete falsification test is to compare the green masks from the fixed latent threshold against independent references—fluorescence reporter signal, manual annotation, or known bacterial density—on the same droplets across multiple time points, focal planes, illumination settings, and post-dispersal frames. If the agreement degrades in any regime the paper claims to cover, the claim of accurate detection and quantification would be falsified. A simpler check is to shift illumination slightly between otherwise identical droplets and ask whether the same manual threshold still separates known sterile crops from biofilm-containing crops.

Watch

Extended reading notes

Core claim

On its own terms, the paper reports that the encoder of a VAE, used without its decoder, is sufficient to segment complex bright-field biofilm images. From two training images per optical setup, 10,000 rotation-invariant radial crops are encoded into a one-dimensional latent variable; a histogram of those values is visually inspected and a threshold is chosen where crops transition from empty background to bacterial content. Mapping each latent value back to its grid position in the droplet produces green/magenta overlays whose green masks quantify aggregates, biofilms, and patches, with aggregate/biofilm distinguished by pixel-area cutoffs ($5000\ \mathrm{px}^2$ for biofilm) rather than by learned features. Time-lapse analysis then resolves a reproducible lifecycle: free-swimming bacteria aggregate, merge into biofilms, develop transient patch-like loops, and disperse after roughly 7.5 hours, with some clusters persisting and possibly regrowing. The paper claims this pipeline detects condition-dependent differences, including faster biofilm formation with glycerol and enhanced biofilm formation at low antibiotic concentrations.

Load-bearing premise

The entire measurement collapses if a VAE trained on only two grayscale images per setup produces a latent space in which one manually chosen threshold does not keep separating bacteria from background across all droplets, time points, illuminations, and conditions.

Editorial extensions

If this is right

  • Automated green masks yield time-resolved relative-area curves for aggregates, biofilms, and patches, and define two quantitative transitions: the biofilm formation trigger (peak positive growth rate of biofilm area) and the dispersal trigger (peak positive growth rate of patch area).
  • The same droplet chip can be read out by fluorescence for high-throughput antibiotic screening or by bright-field time-lapse for lifecycle dynamics, with separate VAE models per optical setup.
  • Pen/Strep screening shows the fraction of proliferated droplets and motile-cell fluorescence fall with antibiotic concentration, while endospore fluorescence peaks at sub-inhibitory concentrations, indicating that low-dose antibiotics can promote biofilm formation.
  • Droplets supplemented with 1% glycerol form biofilms faster, more synchronously, and reach larger absolute biofilm areas than unsupplemented controls, with dispersal onset visible by 6 hours.
  • Segmentation becomes unreliable once dispersal begins because swimming bacteria and debris are present, so the bright-field analysis intentionally covers formation up to dispersal onset; the paper leaves open whether late dark clusters are true regrown biofilms or debris.

Reading between the lines

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

  • The paper leaves implicit that its manual latent threshold could be automated: fitting a two-component mixture model to the latent histogram and comparing the resulting masks with the manually chosen ones would make the pipeline fully unsupervised and more portable across batches.
  • Because each setup is trained on only two images, the threshold's stability under illumination drift, focus drift, or chip-to-chip variation is an open question; a stress test across independently prepared chips and time points would establish whether retraining per setup is always required.
  • The one-dimensional latent space necessarily collapses structural variety, so expanding the latent dimensionality, as the paper itself suggests, could separate endospore-rich zones, aligned cell chains, or different species within the same droplets, connecting bright-field segmentation to the fluorescence readout.
  • The post-dispersal limitation could be addressed by adding temporal tracking of patches and aggregates rather than thresholding each frame independently, which would extend quantitative lifecycle analysis beyond the current 8 to 17.5 hour windows.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The manuscript reports a droplet-based microfluidic platform for time-resolved study of Bacillus subtilis biofilm formation, combined with a Variational Autoencoder (VAE) for unsupervised image segmentation. The VAE is trained on two grayscale images per microscopy setup, and a manually selected threshold in the one-dimensional latent space is used to classify image regions as bacterial structures versus background. From the resulting masks, the authors compute relative areas of aggregates, biofilms, and patches over time, claiming accurate and precise quantification of biofilm growth and dispersal. The platform is also applied to fluorescence-based antibiotic screening and to a glycerol-promotion experiment, with conclusions about the timing of biofilm formation and dispersal.

Significance. If the central claim of accurate, automated quantification were properly validated, the combination of droplet microfluidics with latent-space segmentation would offer a practical high-throughput tool for biofilm research. The microfluidic platform itself—pressure-controlled droplet generation, trapping, multi-injection valve, and compatibility with both fluorescence and bright-field microscopy—is a useful engineering contribution. The idea of bypassing the decoder and thresholding the latent representation is interesting and potentially efficient. However, the paper does not supply any independent validation of the segmentation: there is no comparison to manual annotation, no held-out test set, no sensitivity analysis of the manually chosen threshold, and no justification of the pixel-area cutoffs that define biological categories. As a result, the quantitative biological conclusions are currently not supported by the evidence presented.

major comments (3)
  1. [§2.5, Fig. 1C-D; Abstract; §3.2] The entire segmentation pipeline rests on a single scalar threshold in the latent dimension that is described as 'manually selected' after visual inspection of crops. The VAE is trained on only two grayscale images per setup, and the same datasets used for training and threshold selection are then used to produce all reported area measurements. There is no independent ground truth (e.g., manual segmentation by an expert), no held-out validation, and no sensitivity analysis showing how the measured areas change as the threshold is varied. Since the abstract claims 'accurate detection and quantification' and all downstream conclusions in §3.2 and §3.3 depend on these masks, the central claim of the paper is not supported. The authors should provide at least one of the following: a quantitative comparison against manual annotations, a threshold-sensitivity analysis, or a test on an independent dataset with known ground truth.
  2. [§3.2, classification paragraph] The structural categories 'aggregate', 'biofilm', and 'patch' are defined by arbitrary pixel-area thresholds (300–4999 px², ≥5000 px², and >500 px² inside a biofilm, respectively). These cutoffs are not justified by any biological reference, and the biological interpretations—such as the 'biofilm formation trigger' and 'dispersal trigger' defined from derivatives of area ratios—are read back from measurements that are directly determined by these cutoffs. For example, a cluster crossing the 5000-px² boundary is automatically reclassified from aggregate to biofilm, so the reported transition times partly reflect the chosen thresholds rather than an independent biological property. The authors should demonstrate that their conclusions are robust to reasonable variations of these cutoffs, or validate the cutoffs against an independent measure of biofilm identity (e.g., ECM staining or confocal imaging).
  3. [§3.3 vs §3.2] Section 3.3 explicitly states that after dispersal 'potential biofilm segmentation became increasingly challenging for the current model' and therefore excludes post-dispersal data from the glycerol analysis. Yet Section 3.2 uses the same VAE-based segmentation to quantify the dispersal phase and defines a 'dispersal trigger' from the patch-area curve. This is an internal inconsistency: the model is acknowledged to fail in the very regime where the paper draws key conclusions about dispersal dynamics. The authors should either provide evidence that the segmentation remains reliable during dispersal (e.g., validation against manual annotations in that time window) or restrict the dispersal-related claims accordingly.
minor comments (5)
  1. [Fig. 2A caption] The caption contains a doubled period: 'presence of antibiotics in the medium..' should be corrected.
  2. [PACS/MSC fields] The PACS and MSC codes are given as placeholder values '0000, 1111'; these should be replaced with actual classification codes or omitted.
  3. [Throughout] The term 'V AE' is inconsistently spaced (e.g., 'V AEs' in the Introduction, 'V AE-based' in the Conclusions). Please standardize to 'VAE'.
  4. [§2.5] The description of the random crops as having 'a defined radius of 12 pixels' is ambiguous: it is unclear whether the crops are circular regions of radius 12 pixels or square patches of side 12 pixels, and how this relates to the stated image resolution of 256×256 pixels.
  5. [§3.2, Fig. 3G] The definition of the metric shown in the top panel of Fig. 3G as 'Mean of the positive maximum rate of change' is not clearly explained; please specify how the derivative is computed and how the mean is taken across droplets or time points.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: the VAE-threshold-mask pipeline is a transparent measurement chain, and the paper's only self-citations are non-load-bearing references to a textbook and a software framework.

full rationale

The paper's claimed derivation chain is: VAE encoder maps bright-field crops to a one-dimensional latent variable; a manually selected threshold separates background from bacterial structures; thresholded grid overlays produce green masks; pixel-area cutoffs categorize aggregates, biofilms, and patches; and the relative area time series support the life-cycle and glycerol conclusions. None of these steps is equivalent to its own inputs by construction. The latent threshold is explicitly described as a manually selected parameter, chosen by visual inspection of representative latent-bin crops, and the subsequent area metrics are downstream measurements rather than predictions fitted to those metrics. The reported curves could be biased by threshold drift, illumination changes, or unvalidated area cutoffs, but that is an accuracy and robustness concern, not circularity: the biological conclusions are not used to define or tune the segmentation parameters. The only self-citations are to a Deep Learning textbook [22] and to the Deeplay software framework [26], both by overlapping authors; neither is used as a load-bearing scientific premise, a uniqueness argument, or a justification of the core segmentation claim. Therefore the central derivation is self-contained with respect to circularity, and the score reflects only the presence of minor, non-load-bearing self-citations.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The central quantification rests on a manually chosen latent threshold and hand-set pixel-area cutoffs. The VAE itself is a standard tool; the free parameters are the classification thresholds, not the network weights. No new physical entities are introduced.

free parameters (4)
  • Latent-space threshold separating background from bacterial structures = not reported; manually selected from histogram
    Selected by visual inspection of crops in Section 2.5; the threshold defines which pixels are counted as bacteria, so all area measurements depend on it.
  • Aggregate minimum pixel area = 300 px^2 (3.32 um^2)
    Ad hoc threshold in Section 3.2; regions below this are excluded, affecting aggregate counts and areas.
  • Biofilm minimum pixel area = 5000 px^2 (55.4 um^2)
    Ad hoc cutoff distinguishing aggregates from biofilms in Section 3.2; the reported biofilm dynamics depend on this value.
  • Patch minimum pixel area = 500 px^2 (5.54 um^2)
    Ad hoc threshold for holes within biofilms in Section 3.2; patches with smaller area are discarded.
assumptions (5)
  • domain assumption A scalar threshold in the 1D latent space separates background from bacterial content.
    The VAE is trained on two images per setup and no clustering validation is shown. Section 2.5.
  • domain assumption Two training images capture the full structural variability of the dataset.
    One biofilm and one dispersed image per setup are used; all later images are assumed to lie in the same latent distribution. Section 2.5.
  • domain assumption Bright-field pixel patterns correspond monotonically to bacterial biomass and biofilm structures.
    No fluorescent marker, staining, or confocal verification links the masked green regions to actual cells. Sections 2.5 and 3.2.
  • ad hoc to paper Pixel-area cutoffs correspond to biological categories (aggregate, biofilm, patch).
    Defined without calibration in Section 3.2.
  • domain assumption The water-oil interface in the droplet triggers pellicle biofilm formation.
    The authors cite SI Section 4 for oxygen permeability; in the main text this is assumed for the droplet model.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Latent Space-Driven Quantification of Biofilm Formation using Time Resolved Droplet Microfluidics." pith.science (2026). https://pith.science/paper/ENJTY2MM

@misc{pith2026250707632,
  author       = {Pith},
  title        = {Pith review of: Latent Space-Driven Quantification of Biofilm Formation using Time Resolved Droplet Microfluidics},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ENJTY2MM}},
  note         = {Machine review of arXiv:2507.07632}
}
read the original abstract

Bacterial biofilms play a significant role in various fields that impact our daily lives, from detrimental public health hazards to beneficial applications in bioremediation, biodegradation, and wastewater treatment. However, high-resolution tools for studying their dynamic responses to environmental changes and collective cellular behavior remain scarce. To characterize and quantify biofilm development, we present a droplet-based microfluidic platform combined with an image analysis tool for in-situ studies. In this setup, Bacillus subtilis was inoculated in liquid Lysogeny Broth microdroplets, and biofilm formation was examined within emulsions at the water-oil interface. Bacteria were encapsulated in droplets, which were then trapped in compartments, allowing continuous optical access throughout biofilm formation. Droplets, each forming a distinct microenvironment, were generated at high throughput using flow-controlled pressure pumps, ensuring monodispersity. A microfluidic multi-injection valve enabled rapid switching of encapsulation conditions without disrupting droplet generation, allowing side-by-side comparison. Our platform supports fluorescence microscopy imaging and quantitative analysis of droplet content, along with time-lapse bright-field microscopy for dynamic observations. To process high-throughput, complex data, we integrated an automated, unsupervised image analysis tool based on a Variational Autoencoder (VAE). This AI-driven approach efficiently captured biofilm structures in a latent space, enabling detailed pattern recognition and analysis. Our results demonstrate the accurate detection and quantification of biofilms using thresholding and masking applied to latent space representations, enabling the precise measurement of biofilm and aggregate areas.

Figures

Figures reproduced from arXiv: 2507.07632 by the authors.

Figure 1
Figure 1. Droplet analysis pipeline and latent space visualization of biofilm images. A. In [PITH_FULL_IMAGE:figures/full_fig_p026_1.png] view at source ↗
Figure 2
Figure 2. Biofilm formation in the presence of antibiotics. A. Droplets are considered empty [PITH_FULL_IMAGE:figures/full_fig_p027_2.png] view at source ↗
Figure 3
Figure 3. Spatiotemporal analysis of biofilm formation. A. Bright-field time-lapse images [PITH_FULL_IMAGE:figures/full_fig_p028_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Influence of glycerol on biofilm formation A. Representative images of biofilm for [PITH_FULL_IMAGE:figures/full_fig_p029_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

36 extracted references · 33 canonical work pages

  1. [1]

    Popat, D

    R. Popat, D. M. Cornforth, L. McNally, S. P. Brown, Collective sensing and collective responses in quorum-sensing bacteria, Journal of The Royal Society Interface 12 (2015) 20140882. doi:10.1098/rsif.2014.0882. URL https://royalsocietypublishing.org/doi/10.1098/rsif.2014.0882

    work page arXiv 2015

  2. [2]

    J. G. Mitchell, K. Kogure, Bacterial motility: links to the environment and a driving force for microbial physics., FEMS microbiology ecology 55 (2006) 3–16. doi:10.1111/j.1574-6941.2005.00003.x. URL http://www.ncbi.nlm.nih.gov/pubmed/16420610

    work page arXiv 2006

  3. [3]

    Wadhwa, H

    N. Wadhwa, H. C. Berg, Bacterial motility: machinery and mechanisms., Nature reviews. Microbiology 20 (2022) 161–173. doi:10.1038/s41579- 021-00626-4. URL http://www.ncbi.nlm.nih.gov/pubmed/34548639 18

    work page arXiv 2022

  4. [4]

    M. B. Miller, B. L. Bassler, Quorum sensing in bacteria, Annual Review of Microbiology 55 (2001) 165–199. doi:10.1146/annurev.micro.55.1.165. URL https://www.annualreviews.org/doi/10.1146/annurev.micro.55.1.165

    work page doi:10.1146/annurev.micro.55.1.165 2001

  5. [5]

    M. R. Parsek, E. Greenberg, Sociomicrobiology: the connections between quorum sensing and biofilms, Trends in Microbiology 13 (2005) 27–33. doi:10.1016/j.tim.2004.11.007. URL https://linkinghub.elsevier.com/retrieve/pii/S0966842X04002616

    work page doi:10.1016/j.tim.2004.11.007 2005

  6. [6]

    Arnaouteli, N

    S. Arnaouteli, N. C. Bamford, N. R. Stanley-Wall, ´Akos T. Kov´ acs, Bacillus subtilis biofilm formation and social interactions, Nature Re- views Microbiology 19 (2021) 600–614. doi:10.1038/s41579-021-00540-9. URL https://www.nature.com/articles/s41579-021-00540-9

    work page doi:10.1038/s41579-021-00540-9 2021

  7. [7]

    B. R. Wucher, M. Elsayed, J. S. Adelman, D. E. Kadouri, C. D. Nadell, Bacterial predation transforms the landscape and commu- nity assembly of biofilms, Current Biology 31 (2021) 2643–2651.e3. doi:10.1016/j.cub.2021.03.036

    work page doi:10.1016/j.cub.2021.03.036 2021

  8. [8]

    Espinal, S

    P. Espinal, S. Mart ´ ı, J. Vila, Effect of biofilm formation on the survival of acinetobacter baumannii on dry surfaces, Journal of Hospital Infection 80 (2012) 56–60. doi:10.1016/j.jhin.2011.08.013. URL https://linkinghub.elsevier.com/retrieve/pii/S0195670111003458

    work page doi:10.1016/j.jhin.2011.08.013 2012

Show all 36 references
  1. [9]

    Hathroubi, M

    S. Hathroubi, M. A. Mekni, P. Domenico, D. Nguyen, M. Jacques, Biofilms: Microbial shelters against antibiotics, Microbial Drug Resis- tance 23 (2017) 147–156. doi:10.1089/mdr.2016.0087. URL http://www.liebertpub.com/doi/10.1089/mdr.2016.0087

    2017

  2. [10]

    P´ erez, F

    J. P´ erez, F. J. Contreras-Moreno, F. J. Marcos-Torres, A. Moraleda- Mu˜ noz, J. Mu˜ noz-Dorado, The antibiotic crisis: How bacterial predators can help, Computational and Structural Biotechnology Journal 18 (2020) 2547–2555. doi:10.1016/j.csbj.2020.09.010. URL https://linkin...

    2020 doi

  3. [11]

    D. Sun, K. Jeannot, Y. Xiao, C. W. Knapp, Editorial: Horizontal gene transfer mediated bacterial antibiotic resistance, Frontiers in Microbiology 10 (8 2019). doi:10.3389/fmicb.2019.01933. URL https://www.frontiersin.org/article/10.3389/fmicb.2019.01933/full 19

    2019

  4. [12]

    Kolpen, K

    M. Kolpen, K. N. Kragh, J. B. Enciso, D. Faurholt-Jepsen, B. Lin- degaard, G. B. Egelund, A. V. Jensen, P. Ravn, I. H. M. Math- iesen, A. G. Gheorge, F. B. Hertz, T. Qvist, M. Whiteley, P. Østrup Jensen, T. Bjarnsholt, Bacterial biofilms predominate in both acute and chronic h...

    2022 doi

  5. [13]

    J. M. Sonawane, A. K. Rai, M. Sharma, M. Tripathi, R. Prasad, Microbial biofilms: Recent advances and progress in environmen- tal bioremediation, Science of the Total Environment 824 (6 2022). doi:10.1016/j.scitotenv.2022.153843

    2022

  6. [14]

    C. J. Linton, A. Sherriff, M. R. Millar, Use of a modified robbins device to directly compare the adhesion of staphylococcus epidermidis rp62a to surfaces, Journal of Applied Microbiology 86 (1999) 194–202. doi:10.1046/j.1365-2672.1999.00650.x. URL https://academic.oup.com/jam...

    1999

  7. [15]

    Schwartz, R

    K. Schwartz, R. Stephenson, M. Hernandez, N. Jambang, B. Boles, The use of drip flow and rotating disk reactors for ¡em¿staphylococcus au- reus¡/em¿ biofilm analysis, Journal of Visualized Experiments (12 2010). doi:10.3791/2470. URL https://app.jove.com/t/2470

    2010 doi

  8. [16]

    Azeredo, N

    J. Azeredo, N. F. Azevedo, R. Briandet, N. Cerca, T. Coenye, A. R. Costa, M. Desvaux, G. D. Bonaventura, M. H´ ebraud, Z. Jaglic, M. Kaˇ c´ aniov´ a, S. Knøchel, A. Louren¸ co, F. Mergulh˜ ao, R. L. Meyer, G. Nychas, M. Sim˜ oes, O. Tresse, C. Sternberg, Critical review on bio...

    2017

  9. [17]

    T. S. Kaminski, O. Scheler, P. Garstecki, Droplet microfluidics for microbiology: techniques, applications and challenges, Lab on a Chip 16 (2016) 2168–2187. doi:10.1039/C6LC00367B. URL http://dx.doi.org/10.1039/C6LC00367B http://xlink.rsc.org/?DOI=C6LC00367B

    2016 doi

  10. [18]

    A. G. Niculescu, C. Chircov, A. C. B ˆ ırc˘ a, A. M. Grumezescu, Fab- 20 rication and applications of microfluidic devices: A review (2 2021). doi:10.3390/ijms22042011

    2021 doi

  11. [19]

    Z. Chen, S. Kheiri, E. W. K. Young, E. Kumacheva, Trends in droplet microfluidics: From droplet generation to biomedical applications, Lang- muir 38 (2022) 6233–6248. doi:10.1021/acs.langmuir.2c00491. URL https://pubs.acs.org/doi/10.1021/acs.langmuir.2c00491

    2022 doi

  12. [20]

    Jiang, H.-Z

    M.-Z. Jiang, H.-Z. Zhu, N. Zhou, C. Liu, C.-Y. Jiang, Y. Wang, S.-J. Liu, Droplet microfluidics-based high-throughput bacterial cultivation for validation of taxon pairs in microbial co-occurrence networks, Scien- tific Reports 12 (1) (2022) 18145. doi:10.1038/s41598-022-23000...

    2022 doi

  13. [21]

    D. P. Kingma, M. Welling, Auto-encoding variational bayes, in: Pro- ceedings of the 2nd International Conference on Learning Representa- tions (ICLR), 2014, arXiv:1312.6114

    2014 arXiv

  14. [22]

    Volpe, B

    G. Volpe, B. Midtvedt, J. Pineda, H. K. Moberg, H. Bachimanchi, J. B. Pereira, C. Manzo, Deep Learning Crash Course, No Starch Press, 2025

    2025

  15. [23]

    Gu, Q.-L

    H.-J. Gu, Q.-L. Sun, J.-C. Luo, J. Zhang, L. Sun, A first study of the virulence potential of a bacillus subtilis isolate from deep-sea hydrothermal vent, Frontiers in Cellular and Infection Microbiology 9 (5 2019). doi:10.3389/fcimb.2019.00183. URL https://www.frontiersin.org...

    2019

  16. [24]

    H. Liu, L. Nan, F. Chen, Y. Zhao, Y. Zhao, Functions and applications of artificial intelligence in droplet microfluidics, Lab on a Chip 23 (2023) 2497–2513. doi:10.1039/D3LC00224A. URL http://xlink.rsc.org/?DOI=D3LC00224A

    2023 doi

  17. [25]

    C. Feng, K. Takahashi, J. Zhu, Simple one-step and rapid patterning of pdms microfluidic device wettability for pdms shell produc- tion, Frontiers in Bioengineering and Biotechnology 10 (4 2022). doi:10.3389/fbioe.2022.891213. URL https://www.frontiersin.org/articles/10.3389/f...

    2022

  18. [26]

    Midtvedt, J

    B. Midtvedt, J. Pineda, H. Klein Morberg, H. Bachimanchi, M. Gran- fors, A. Lech, C. Manzo, G. Volpe, Deeplay. URL https://github.com/DeepTrackAI/deeplay 21

  19. [27]

    Guilhen, C

    C. Guilhen, C. Forestier, D. Balestrino, Biofilm dispersal: multiple elab- orate strategies for dissemination of bacteria with unique properties, Molecular Microbiology 105 (2017) 188–210. doi:10.1111/mmi.13698. URL https://onlinelibrary.wiley.com/doi/10.1111/mmi.13698

    2017 doi

  20. [28]

    ´Akos T Kov´ acs, N. R. Stanley-Wall, Biofilm dispersal for spore release in bacillus subtilis., Journal of bacteriology 203 (2021) e0019221. doi:10.1128/JB.00192-21. URL http://www.ncbi.nlm.nih.gov/pubmed/33927051 http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=PMC8223919

    2021

  21. [29]

    Bartolini, S

    M. Bartolini, S. Cogliati, D. Vileta, C. Bauman, L. Rateni, C. Le˜ nini, F. Arga˜ naraz, M. Francisco, J. M. Villalba, L. Steil, U. V¨ olker, R. Grau, Regulation of biofilm aging and dispersal in bacillus subtilis by the alternative sigma factor sigb, Journal of Bacteriology 2...

    2019 doi

  22. [30]

    Nishikawa, K

    M. Nishikawa, K. Kobayashi, Calcium prevents biofilm disper- sion in bacillus subtilis, Journal of Bacteriology 203 (6 2021). doi:10.1128/JB.00114-21. URL https://journals.asm.org/doi/10.1128/JB.00114-21

    2021 doi

  23. [31]

    R. Zhao, Y. Song, Q. Dai, Y. Kang, J. Pan, L. Zhu, L. Zhang, Y. Wang, X. Shen, A starvation-induced regulator, rovm, acts as a switch for planktonic/biofilm state transition in yersinia pseudotuberculosis, Scientific Reports 7 (2017) 639. doi:10.1038/s41598-017-00534-9. URL ht...

    2017 doi

  24. [32]

    Gjermansen, P

    M. Gjermansen, P. Ragas, C. Sternberg, S. Molin, T. Tolker-Nielsen, Characterization of starvation-induced dispersion in pseudomonas putida biofilms, Environmental Microbiology 7 (2005) 894–904. doi:10.1111/j.1462-2920.2005.00775.x. URL https://onlinelibrary.wiley.com/doi/10.1...

    2005

  25. [33]

    Harmsen, L

    M. Harmsen, L. Yang, S. J. Pamp, T. Tolker-Nielsen, An update on pseudomonas aeruginosa biofilm formation, tolerance, and dispersal, FEMS Immunology & Medical Microbiology 59 (2010) 253–268. 22 doi:10.1111/j.1574-695X.2010.00690.x. URL https://academic.oup.com/femspd/article-l...

    2010 arXiv

  26. [34]

    Zhang, A

    W. Zhang, A. Seminara, M. Suaris, M. P. Brenner, D. A. Weitz, T. E. Angelini, Nutrient depletion in bacillus subtilis biofilms trig- gers matrix production, New Journal of Physics 16 (2014) 015028. doi:10.1088/1367-2630/16/1/015028. URL https://iopscience.iop.org/article/10.10...

    2014 doi

  27. [35]

    Hobley, S

    L. Hobley, S. Kim, Y. Maezato, S. Wyllie, A. Fairlamb, N. Stanley-Wall, A. Michael, Norspermidine is not a self-produced trigger for biofilm disassembly, Cell 156 (2014) 844–854. doi:10.1016/j.cell.2014.01.012. URL https://linkinghub.elsevier.com/retrieve/pii/S0092867414000233

    2014 doi

  28. [36]

    Kim, J.-H

    S.-K. Kim, J.-H. Lee, Biofilm dispersion in pseudomonas aeruginosa, Journal of Microbiology 54 (2016) 71–85. doi:10.1007/s12275-016-5528- 7. URL http://link.springer.com/10.1007/s12275-016-5528-7 23 Variational Autoencoder Encoder Latent space Input 42.4 μm A. B. Bacteria in s...

    2016 doi

Pith tools

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