Automated 3D Optical Coherence Tomography to Elucidate Biofilm Morphogenesis Over Large Spatial Scales
August 21st, 2019
Microbial biofilms form complex architectures at interphases and develop into highly scale-dependent spatial patterns. Here, we introduce an experimental system (hard- and software) for the automated acquisition of 3D optical coherence tomography (OCT) datasets. This toolset allows the non-invasive and multi-scale characterization of biofilm morphogenesis in space and time.
Here we present an automated system based on optical coherence tomography or OCT. This allows us to monitor biofilms'structure over large spatial scales or extended periods of time. OCT imaging is well suited to resolve structures in the micrometer range, however it's currently limited to a maximum area of about 250 square millimeters.
Biofilms often exceed this scale, especially when the differentiation is driven by large-scale environmental gradients. The experimental setup allow us to monitor the three-dimensional morphogenesis of biofilms over large spatial scales and extended periods of time. The system is fast, precise, and it works autonomously.
We studied the morphogenesis of biofilms in streams where they drive important ecosystem processes. However, the system may be used to study biofilms in other natural system or engineered environments. The software for positioning, image acquisition, and analysis is written in Python.
They are available through Jupyter Notebooks. These are user-friendly, freely available, and flexible solutions. We believe that the visual representation of the setup helps other users to reproduce the installation and to better understand the software.
We hope this inspires other researchers to adopt similar approaches. Here is an overview of the installation. The system is composed of a precision positioning device, the OCT probe, and it is assembled around a plexiglass flume.
Begin by wiring the positioning device by following along with instructions posted on GitHub. Once connected, install the GRBL server as described in a separate GitLab page. Positioning system can now be controlled through this web page;alternatively it can be controlled through a Python script as shown in the worked example.
Position the computer and the OCT base unit on a bench next to the experimental setup containing microfluidic devices, flow chambers, flumes, or filtration systems. If not already installed, install the OCT system together with the available software as described by the manufacturer. Then install the software packages for automated OCT scan acquisition as described in the GitLab documents linked to here.
To begin image acquisition, mount the optical coherence tomography probe to the positioning device using a compatible dovetail holder. If required, install an immersion adapter on the objective lens, then power on the OCT system and the positioning device. Open the commercial OCT software, locate a site of interest, focus on the sample, and adjust the reference arm and light source intensity for optimal image quality.
Note the coordinates and repeat this procedure for a number of positions while maintaining the same reference arm length and intensity. Open the ImageAcquisition. ipynb file found in this article's Supplementary File 2 in Jupyter Notebook.
Each cell contains code to perform specific tasks and can be run separately via pressing Cell and then Run or Control and Enter or Shift and Enter. Follow the worked example to set the path to the required libraries to connect the positioning device to calibrate the positioning device to initialize the OCT scanner. Then adjust the acquisition parameters, including the refractive index, the size of the field of view, and the number of A scans per B scan.
Further, set the signal boundaries of the OCT scan based on intensity histograms of preliminary scans and the destination folder for acquired data and metadata. Depending on the field of view and resolution, the file size may reach up to 1.5 gigabytes per OCT scan. These two parameters determine the size of the voxels in the final data set and the size of the output file.
They should match the optical resolution of the OCT probe. As highlighted in the worked example, you may acquire a single OCT scan with default parameters or acquire a single scan specifying a different set of parameters. You may also provide specific coordinates to move the positioning device and acquire a single OCT scan.
This feature allows you to repeatedly return to the exact same position in the experiment with high spatial accuracy. Data is saved in 8bit. raw format to save storage space.
Metadata, including the OCT settings and coordinates, are saved in the same folder in a json file with the same naming convention. Alternatively, specify a list of positions of interest and acquire the respective OCT scans automatically. In order to characterize biofilm morphological structures across large environmental gradients, acquire scans in a mosaic pattern.
For this, specify the number of neighboring tiles with a default overlap of 30%The raw OCT scans appear distorted. This is due to differences in path length through the optical system. We developed an algorithm which correct this distortion as shown in the worked examples.
To begin image correction, open the Jupyter notebook ImageProcessing.ipynb. Following this example, first crop the OCT scans in order to exclude spurious signals and reorient the data set so that the biofilm appears above the substratum. Next correct for spherical aberration.
To accomplish this, the algorithm localizes a highly reflective flat surface in the OCT scan and uses this as reference to flatten the scans. Across a 20 by 20 grid, the algorithm then identifies local maxima in signal intensity to localize the reference surface. Then a second order polynomial surface is fitted across these points and used to shift each pixel of the OCT scan in Z direction.
The parameters of this algorithm should be adjusted to the characteristics of the OCT scan. This correction enables a homogeneous reference surface across multiple images and thus facilitates stitching of large-scale images. Once the image has been flattened, the images are corrected for background noise by identifying an empty area of the image above the biofilm and subtracting average background intensity.
Next compute an elevation map from the 3D OCT data set. In order to do so, define a reference surface such as the substratum and choose an appropriate intensity threshold. Then an elevation map is rendered with the height of the biofilm reported as grayscale value.
If images were acquired in a mosaic pattern, stitch the respective elevation maps by applying the stitching algorithm. Using automated OCT imaging, the spatiotemporal morphogenesis of phototrophic stream biofilms was examined using flume experiments. The flumes are made of plexiglass and gradually widen from the in to the outflow.
This results in a gradient in flow velocity. Here is an elevation map of a biofilm growing along the entire flow velocity gradient. Importantly, the automated OCT imaging system allows a continuous measurement of structural parameters such as biofilm thickness, roughness, and biovolume under differing flow conditions ranging from low-flow velocity to high-flow velocity conditions.
Along with morphological changes, average biovolume significantly decreased as a function of distance from the inlet in the flume. The quality of the OCT scans critically depends on the reference arm length and the focus distance. You may need to readjust this parameter during the experiments.
To ensure the accuracy of the positioning device, remember to perform regularly homing operations. This automated imaging device can be readily coupled with microsensor profiling for a functional characterization of biofilms. OCT is an emerging imaging technique and we anticipate that the system presented here stimulates research on biofilm structure.
This may be relevant for technologies such as drinking water treatment or bioprocessing.
View the full transcript and gain access to thousands of scientific videos
Overview
This article presents an automated system utilizing optical coherence tomography (OCT) to monitor the structure of microbial biofilms. The system enables the non-invasive observation of biofilm morphogenesis over large spatial scales and extended periods of time.
Key Study Components
Area of Science
- Microbial biofilms
- Optical coherence tomography
- Biofilm morphogenesis
Background
- Microbial biofilms exhibit complex architectures.
- Biofilms can develop spatial patterns influenced by environmental gradients.
- Current imaging techniques have limitations in spatial coverage.
- OCT is effective for resolving structures in the micrometer range.
Purpose of Study
- To develop a system for automated OCT imaging of biofilms.
- To monitor biofilm development over time and space.
- To enhance the understanding of biofilm morphogenesis.
Methods Used
- Automated optical coherence tomography system.
- Non-invasive imaging technique.
- Monitoring over large spatial scales.
- Extended observation periods.
Main Results
- The system allows for fast and precise imaging of biofilms.
- It can capture three-dimensional morphogenesis.
- Effective for monitoring biofilms exceeding 250 square millimeters.
- Facilitates the study of biofilm differentiation influenced by environmental factors.
Conclusions
- The automated OCT system is a valuable tool for biofilm research.
- It provides insights into the dynamic nature of biofilm structures.
- This approach can advance our understanding of microbial ecology.
Frequently Asked Questions
Executive Industry Relevance
Automated 3D optical coherence tomography enables non-invasive, high-resolution monitoring of biofilm morphogenesis across large spatial and temporal scales, addressing a key limitation in current biofilm imaging. This capability supports mechanistic de-risking in early discovery by providing quantitative structural data under controlled environmental gradients, facilitating target validation and predictive confidence in antimicrobial or anti-biofilm therapeutic development. The system’s automation and scalability enhance reproducibility and translational continuity from discovery through preclinical evaluation.
Strategic Applications in Biopharma R&D
Early Discovery & Target Validation
- Scientific Value: Enables interrogation of biofilm structural responses to environmental stimuli, supporting hypothesis testing of virulence or resistance mechanisms.
- Operational Value: Provides automated, reproducible 3D imaging that reduces variability in morphological assessment across experimental conditions.
- Predictive Value: Generates quantitative metrics such as thickness, roughness, and biovolume to correlate structural changes with functional outcomes in biofilm disruption studies.
Screening & Assay Development
- Scientific Value: Produces standardized, distortion-corrected elevation maps enabling reliable comparison of biofilm architecture across compound or condition screens.
- Operational Value: Integrates with Python-based Jupyter notebooks for streamlined acquisition, processing, and analysis, supporting assay standardization and throughput.
- Scalability: Mosaic stitching capability allows coverage of large biofilm areas exceeding conventional OCT limits, enabling screening under physiologically relevant gradients.
Translational & Preclinical Research
- Translational Continuity: Supports longitudinal monitoring of biofilm development in flow-based systems that mimic natural or engineered environments, relevant to infection or industrial fouling models.
- Mechanistic De-risking: Enables correlation of structural morphogenesis with environmental gradients, informing dose-response or timing interventions in preclinical studies.
- Biomarker Alignment: Structural parameters like biovolume and roughness serve as quantifiable readouts for target engagement or therapeutic efficacy in biofilm models.
Pipeline & Workflow Integration
The automated OCT system fits within the discovery continuum from early target validation through lead identification and preclinical evaluation, particularly for anti-infective or anti-fouling strategies targeting biofilm-forming pathogens.
- Discovery Biology: Facilitates hypothesis-driven structural analysis of biofilm communities under controlled flow and nutrient gradients, clarifying mechanisms of adaptation and resilience.
- Screening: Enables standardized, quantitative imaging of biofilm morphology across compound libraries or genetic variants, supporting hit-to-lead progression.
- Analytics: Outputs elevation maps and structural metrics (thickness, roughness, biovolume) that allow objective comparison of biofilm states across time and treatment conditions.
- Translational Research: Bridges in vitro biofilm models to preclinical relevance by capturing spatiotemporal dynamics under flow, mimicking mucosal or industrial surfaces.
- Enterprise Reuse: Modular design with open-source Python scripts and Jupyter notebooks allows adaptation across lab settings and model systems, promoting platform reuse.
Operational & Enterprise Impact
- Scientific Value: Reduces mechanistic ambiguity in biofilm studies by providing objective, 3D structural data under defined environmental conditions.
- Operational Value: Ensures reproducibility through automated positioning, calibration, and processing workflows, minimizing user-dependent variability.
- Strategic Value: Improves go/no-go decisions in anti-biofilm programs by delivering predictive structural phenotypes linked to functional outcomes.
- Portfolio Impact: Enables risk-adjusted prioritization of candidates based on their ability to disrupt or prevent biofilm formation under relevant physiological gradients.
Implementation Considerations
- Requires expertise in optical imaging, biofilm culture, and Python-based data processing for system setup and workflow integration.
- Depends on commercially available OCT hardware, precision positioning stages, and stable environmental control (e.g., flow chambers or flumes) for longitudinal imaging.
- Necessitates cross-team standardization of imaging protocols, file naming, and metadata handling to ensure data comparability across studies.
- Adaptation to different model systems may require adjustments to immersion optics, scanning parameters, and reference surface correction based on sample opacity and refractive index.
- Practical limitations include file size management (up to 1.5 GB per scan) and the need for regular homing of the positioning device to maintain spatial accuracy over extended runs.
