Excitation-Scanning Hyperspectral Imaging Microscopy to Efficiently Discriminate Fluorescence Signals
August 22nd, 2019
Spectral imaging has become a reliable solution for identification and separation of multiple fluorescence signals in a single sample and can readily distinguish signals of interest from background or autofluorescence. Excitation-scanning hyperspectral imaging improves on this technique by decreasing the necessary image acquisition time while simultaneously increasing the signal-to-noise ratio.
This protocol is significant as it describes a method which can be used to separate several fluorescent signals in a single sample. Compared to traditional methods, excitation-scanning can increase the speed of the experiment and improve the sensitivity of the data. By using different illumination sources such as a supercontinuum laser, this method may be applicable to spinning disk or laser scanning confocal microscope systems.
When performing excitation-scannings, spectral imaging microscopy, it is important to check that the excitation peak of each fluorescent label occurs at a different wavelength. To begin, set up an inverted fluorescence microscope and camera as described in the text protocol. Then load the sample on the stage.
From the drop-down menu of the micromanager main window, select 475 nanometers for initial sample viewing. Then double-click at the box next to exposure and enter 100 to set the exposure time at 100 milliseconds. Finally, click live to view the sample.
Now click on Auto. The automatic intensity viewing range button to bring the minimum and maximum values into meaningful, visual ranges. Next, use the microscope's focus knobs to focus the sample.
Once in focus, adjust the position of the sample so that it's edge is in the center of the field of view. Then fine-tune the focus. The sample is in focus when the edge features in the image appear sharp.
Now, open the multi-dimensional acquisition tool. In the menu, click the load button and choose a preset setting containing an appropriate wavelength range for spectral acquisition. Once the acquisition settings are confirmed, move to a blank region of the sample and click acquire to collect a spectral image stack containing background and noise to use for subtraction later.
Then, move about the sample to find the brightest regions. Samples are likely to have more intense fluorescence away from the edge of the sample. Once the brightest region is located, take a single image stack.
Load the acquired image into Image J and use Image J's measure function to confirm that no wavelengths contain overexposed pixels. If a reduced exposure time is necessary due to overexposure, the background image should be retaken with a new exposure time as well. If any of the images were found to contain the upward detectional limit of the camera, adjust the exposure time of the spectral scan to ensure that the maximum signal throughout the spectral range does not exceed the dynamic range of the camera.
With an appropriate exposure time now determined, place an unlabeled sample onto the stage. Acquire spectral image data from the unlabeled sample to determine if there is any underlying autofluoresence. Then, place a single-labeled sample on the stage.
Acquire spectral image data from each of the single-labeled samples to use as spectral controls to build a spectral library. Perform the scan for each control sample using an identical wavelength range exposure time and camera settings. Next, place a slide containing the experimental sample on the microscope.
Select a field of view that has appropriately labeled cells and acquire the spectral image data from the sample using the determined acquisition settings. Correct images to a flat spectral response by subtracting the background spectrum from the sample image and multiplying the image by the correction coefficient. Here, a MATLAB script is used to quickly process the images.
Next, generate the spectral library. For best results, normalize each end member to the wavelength band of the greatest intensity by determining which wavelength band has the most intense value and dividing the measurement at each wavelength band by that maximum value. Then save the spectral data as a dat file.
When complete, unmix the spectral image data. The unmixing step will generate an abundance image for each fluorescent label, where abundance is the amount of relative fluorescent signal in the image from the prospective label. Next, open each unmixed image to visually inspect the distribution of pure components.
Compare the magnitude of the error image to those of the unmixed images. If the magnitude of the error image is similar to the unmixed abundance images, signals in the spectral image data may not be accurately accounted for. Finally, check that there are no unidentified residual structures by comparing the error image to the unmixed images.
When performed correctly, the spectral unmixing process allows separation of a spectral image stack into respective contributions from each fluorescence label. Here, an unmixed spectral image set with individual images highlighting airway smooth muscle cell autofluorescence, a GCaMP probe and a mitochondrial label. There is high error associated with the nuclear and perinuclear regions of the cells indicating that the measured spectra of those regions are not well accounted for by the spectra in the spectral library.
Pure spectra collected from labeled cells that are also autofluorescent will likely contaminate the spectral profile for the pure component. Here, one can see the difference of unmixing with pure spectra contaminated with autofluorescence, versus the corrected pure spectra after scaled subtraction. Proper data acquisition is essential when imaging sensitive samples as they may only survive a single imaging session unlike the background and controls which may be reimaged.
Ensure that the proper hardware and acquisition settings are predetermined as much as possible. As a next step, videos created from time-lapse data acquired using this system can be used to track fluorescent sources in discreet locations over time such as second messenger signal localization. Take care when using powerful light sources to avoid eye exposure.
Also, be sure to wear proper gloves when handling biological labels.
View the full transcript and gain access to thousands of scientific videos
Overview
This protocol describes a method for separating multiple fluorescent signals in a single sample using excitation-scanning hyperspectral imaging. This technique enhances the speed of experiments and improves data sensitivity compared to traditional methods.
Key Study Components
Area of Science
- Fluorescence microscopy
- Imaging techniques
- Signal processing
Background
- Fluorescence signals can be difficult to distinguish due to background noise.
- Traditional imaging methods may be slow and less sensitive.
- Excitation-scanning hyperspectral imaging offers a solution to these challenges.
- Different illumination sources can be utilized for improved results.
Purpose of Study
- To provide a reliable method for separating multiple fluorescence signals.
- To increase the speed and sensitivity of fluorescence imaging experiments.
- To demonstrate the applicability of this method to various microscopy systems.
Methods Used
- Setup of an inverted fluorescence microscope and camera.
- Use of a supercontinuum laser as an illumination source.
- Loading of samples onto the microscope stage.
- Verification of excitation peaks at different wavelengths for each fluorescent label.
Main Results
- Excitation-scanning hyperspectral imaging significantly reduces acquisition time.
- Improved signal-to-noise ratio enhances data quality.
- Method is adaptable to spinning disk or laser scanning confocal systems.
- Successful separation of multiple fluorescence signals demonstrated.
Conclusions
- Excitation-scanning hyperspectral imaging is a valuable tool for researchers.
- This method can lead to more efficient and sensitive fluorescence imaging.
- Future applications may expand to various fields of biological research.
Frequently Asked Questions
Executive Industry Relevance
Excitation-scanning hyperspectral imaging addresses a key bottleneck in fluorescence-based discovery workflows: the inability to rapidly and sensitively resolve overlapping signals from multiple probes in complex biological samples. By improving acquisition speed and signal-to-noise ratio, this method enhances predictive confidence in target validation and mechanistic de-risking, particularly in early-stage phenotypic screening where signal separation is critical. Its compatibility with standard microscopy platforms supports scalable integration into enterprise imaging pipelines.
Strategic Applications in Biopharma R&D
Early Discovery & Target Validation
- Scientific Value: Enables interrogation of therapeutic hypotheses by isolating fluorescence signals from distinct probes, clarifying pathway engagement and target specificity.
- Operational Value: Reduces mechanistic ambiguity through spectral unmixing, supporting functional target validation with quantitative signal deconvolution.
- Predictive Value: Improves confidence in lead identification by minimizing false positives from spectral overlap or autofluorescence interference.
Screening & Assay Development
- Assay Readiness: Prepares validated biological systems for downstream screening by generating spectral libraries that ensure reproducible, quantitative readouts.
- Scalability: Supports platform reuse across projects through standardized acquisition and unmixing workflows, reducing assay redevelopment time.
- Quantitative Output: Generates abundance images per fluorophore, enabling precise comparison of signal intensity across conditions for hit selection.
Translational & Preclinical Research
- Translational Continuity: Maintains signal fidelity from discovery through preclinical models by correcting for autofluorescence and spectral contamination.
- Risk-Adjusted Advancement: Supports go/no-go decisions by providing clean, unmixed channel data that reflect true biological signal rather than artifact.
- Mechanistic De-risking: Isolates signals from organelles or compartments (e.g., mitochondria, cytosol) to validate target engagement in disease-relevant systems.
Pipeline & Workflow Integration
This method fits within the discovery continuum from early target validation through lead optimization, where multiplexed fluorescence readouts are used to assess mechanism, toxicity, and pathway modulation. It enables reliable handoff between biology and analytics teams by delivering standardized, quantifiable imaging outputs.
- Discovery Biology: Supports hypothesis testing and pathway clarification by resolving signals from multiple labels (e.g., GCaMP, mitochondrial dye, autofluorescence) in live-cell imaging.
- Screening: Enhances assay readiness through reproducible spectral acquisition and background subtraction, improving Z-factor and hit-call confidence.
- Analytics: Delivers unmixed abundance maps and error metrics that enable statistical comparison of fluorophore contributions across experimental groups.
- Translational Research: Connects to preclinical validation by ensuring that fluorescence signals used as biomarkers are not confounded by spectral bleed-through or autofluorescence.
- Enterprise Reuse: Establishes a centralized spectral library and SOP for hyperspectral acquisition, promoting cross-project consistency and reducing redundant method development.
Operational & Enterprise Impact
- Scientific Value: Increases predictive confidence in target validation by reducing spectral ambiguity and improving signal fidelity.
- Operational Value: Enhances reproducibility and throughput via faster acquisition and standardized calibration using supercontinuum illumination.
- Strategic Value: Improves capital efficiency by enabling more data per sample and reducing repeat imaging due to poor signal separation.
- Portfolio Impact: Supports risk-adjusted prioritization by delivering cleaner, more interpretable imaging data for decision gates between discovery and preclinical stages.
Implementation Considerations
- Requires expertise in fluorescence microscopy, spectral imaging, and image processing tools such as ImageJ and MATLAB.
- Dependent on tunable illumination sources (e.g., supercontinuum laser) and detectors with sufficient dynamic range and sensitivity.
- Necessitates cross-team standardization of acquisition settings, exposure times, and spectral library generation for reproducible results.
- Must account for autofluorescence in biological samples, requiring control measurements and spectral subtraction to avoid contamination of pure component spectra.
- Limited by phototoxicity risks in live-cell applications, necessitating careful exposure management to preserve sample viability during acquisition.
