How to cite this article: Wu, J. et al. A long Stokes shift red fluorescent Ca2+ indicator protein for two-photon and ratiometric imaging. Nat. Commun. 5:5262 doi: 10.1038/ncomms6262 (2014).
ratiometric imaging image j software 24
Multiplexed bioassays, in which multiple analytes of interest are probed in parallel within a single small volume, have greatly accelerated the pace of biological discovery. Bead-based multiplexed bioassays have many technical advantages, including near solution-phase kinetics, small sample volume requirements, many within-assay replicates to reduce measurement error, and, for some bead materials, the ability to synthesize analytes directly on beads via solid-phase synthesis. To allow bead-based multiplexing, analytes can be synthesized on spectrally encoded beads with a 1:1 linkage between analyte identity and embedded codes. Bead-bound analyte libraries can then be pooled and incubated with a fluorescently-labeled macromolecule of interest, allowing downstream quantification of interactions between the macromolecule and all analytes simultaneously via imaging alone. Extracting quantitative binding data from these images poses several computational image processing challenges, requiring the ability to identify all beads in each image, quantify bound fluorescent material associated with each bead, and determine their embedded spectral code to reveal analyte identities. Here, we present a novel open-source Python software package (the mrbles analysis package) that provides the necessary tools to: (1) find encoded beads in a bright-field microscopy image; (2) quantify bound fluorescent material associated with bead perimeters; (3) identify embedded ratiometric spectral codes within beads; and (4) return data aggregated by embedded code and for each individual bead. We demonstrate the utility of this package by applying it towards analyzing data generated via multiplexed measurement of calcineurin protein binding to MRBLEs (Microspheres with Ratiometric Barcode Lanthanide Encoding) containing known and mutant binding peptide motifs. We anticipate that this flexible package should be applicable to a wide variety of assays, including simple bead or droplet finding analysis, quantification of binding to non-encoded beads, and analysis of multiplexed assays that use ratiometric, spectrally encoded beads.
A) Diagram of the MRBLE library assay. B) Photograph and optics schematic of the imaging setup (left) and overview of data analysis pipeline (right). AL = arc lamp; LE = LED light engine; LLG = liquid light guide; CM = condenser mount; FC = focusing collimator; S = sample and stage; F = UV blocking filter mounted on top of the objective within a 3D printed holder; obj = CFI Plan Achromat 4x objective, NA of 0.2; FT = filter turret; M = mirror; FW = filter wheel; EMCCD = Andor iXon EMCCD camera. C) From left to right: bright-field image of MBRLEs; images of the 9 emission channels for decoding; image of the Cy5 channel used for assay quantification. D) Close-ups of bright-field image of MBRLEs; one emission channel (620 nm, which corresponds to a specific peak in Europium); and image of the Cy5 channel used for assay quantification.
Accessing the large code spaces possible with LNP-based encoding requires collecting sufficient photons from each bead to limit shot noise and accurately discriminate between many different intensity levels. LNP emission lifetimes on the order of μs to ms (more than 1,000-fold slower than fluorescence emission) [20]. As a result, sufficient photons cannot be collected via traditional flow cytometry, in which beads traverse photon detectors over short transit times ( 10 μs). Instead, beads can be imaged via microscopy with imaging times in each LNP channel adjusted to allow discrimination of the required number of LNP intensity levels while maximizing overall throughput. Identifying analytes and measuring binding therefore requires the development of image processing software capable of identifying individual beads from microscopy images, quantifying fluorescent material bound to each bead, and calculating ratios of embedded LNPs to identify the bead code and thus, the identity of the bound analyte (Fig 1B). Identifying bead codes and quantifying binding additionally requires segmentation of each bead into an outer shell and an inner core, as LNPs that comprise the embedded spectral codes are located within the central core, while bound probes produce a ring of fluorescence around the outer bead margins (Fig 1C and 1D). While multiple commercial software packages exist for decoding ratiometric spectral codes (e.g. BD FCAP, Illumina xPONENT, Bio-Rad Bio-Plex), these are optimized for flow cytometry data rather than images and are therefore incompatible with the use of LNPs. In addition, these commercial packages are closed-source, preventing critical modifications required for development of new assays.
Here, we present mrbles, an open-source software package for analysis of images acquired from multiplexed bioassays using spectrally encoded beads. This software is written in Python, a widely-used open-source programming language, and accompanied by an example Jupyter Notebook to be accessible even to users with limited programming experience. While we present this package as a method for decoding MRBLEs, mrbles should be broadly applicable to decoding and quantifying binding to unencoded beads as well as beads embedded with any ratiometric codes, including those based on organic fluorophores or quantum dots. Therefore, the analysis can be used for a wide range of potential applications, from imaging bead-bound libraries incubated with different concentrations of fluorescently-labeled macromolecules to extract binding affinities in high-throughput to imaging bead-bound libraries over time to extract binding kinetics. Unlike commercial software packages, which require images and inputs generated from proprietary commercial hardware, the mrbles package is compatible with a wide variety of inputs.
The mrbles analysis package is designed to automate image analysis and data extraction for multiplexed bioassays and to provide subsequent quality control, data analysis, and visualization tools. The software is set up in a modular fashion to provide data from each step in the pipeline, thereby enhancing generalizability (Fig 2).
In the absence of OME-TIFF hyperstack images (e.g. if users only have a single image for bead finding from bright-field images, or users acquired images with a different acquisition software), the mrbles.Images class can alternately load user-provided multidimensional NumPy arrays in a Python dictionary with an optional list of names for each channel.
Extracting quantitative information from individual beads requires the ability to reliably identify beads, even when packed together, while excluding other small contaminants (e.g. dust particles). Under bright-field illumination, MRBLEs appear as dark rings visible within a bright background (Fig 3A and 3B), allowing bead identification via a simple thresholding procedure in which all pixels that exceed a particular value are used to define bead margins. Unlike traditional circle finding methods (e.g. Hough transform), this preserves morphological information about each MRBLE that can be used in downstream filtering to eliminate unwanted particles. The use of bright-field images to identify beads renders the mrbles package compatible with assays imaged with low-cost bright-field illumination. To ensure that bead finding is robust and reproducible across different microscope configurations, we tested bead finding using bright field images obtained using transmission illumination from above the sample as well as using illumination reflected via a dichroic mirror positioned within a filter turret below the sample. Both sets of sample images are included as supplementary information. The software also includes built-in functionality to enable bead finding via a simple circle finding method (as demonstrated in the example Jupyter notebook).
To calculate LNP ratios associated with each bead, the mrbles.Extract class: (1) identifies all pixels in the linearly unmixed and ratiometric images associated with each bead core, and then (2) calculates median LNP levels and ratios for all of these pixels. To remove objects that resemble beads under bright-field imaging but are not true encoded particles (e.g. air bubbles), mrbles.Extract can filter out particles with invariant LNP levels and unmixed background levels that deviate significantly from the calculated mean value (using a user-defined threshold with a default value of > 2 standard deviations). This strategy ensures that all identified beads used in downstream analysis are actual MRBLEs with an accuracy of > 99.9% [15].
For demonstration and troubleshooting, the software package comes with a set of images from a MRBLEs assay measuring interactions between a fluorescently-labeled (Cy5) protein and peptides on different encoded beads. The package also includes example images for calculating individual LNP reference spectra, and a flat-field image for use in flat-field correcting measured Cy5 intensities. The GitHub repository also provides example Jupyter Notebook files analyzing these images with extensive explanations provided for each pipeline step described in this paper, following the structure of the software depicted in Fig 2. The source code is documented following Python NumPy docstring convention, giving users instant access to information to which parameters to provide and which methods are available. This documentation is also available on the GitHub Pages:
Fluorescence lifetime imaging (FLIM) uses the fact that the fluorescence lifetime of a fluorophore depends on its molecular environment but not on its concentration. Molecular effects in a sample can therefore be investigated independently of the variable, and usually unknown concentration of the fluorophore. There is a variety of technical solutions of lifetime imaging in microscopy. The technical part of this paper focuses on time-domain FLIM by multidimensional time-correlated single photon counting, time-domain FLIM by gated image intensifiers, frequency-domain FLIM by gain-modulated image intensifiers, and frequency-domain FLIM by gain-modulated photomultipliers. The application part describes the most frequent FLIM applications: Measurement of molecular environment parameters, protein-interaction measurements by Förster resonance energy transfer (FRET), and measurements of the metabolic state of cells and tissue via their autofluorescence. Measurements of local environment parameters are based on lifetime changes induced by fluorescence quenching or conformation changes of the fluorophores. The advantage over intensity-based measurements is that no special ratiometric fluorophores are needed. Therefore, a much wider selection of fluorescence markers can be used, and a wider range of cell parameters is accessible. FLIM-FRET measures the change in the decay function of the FRET donor on interaction with an acceptor. FLIM-based FRET measurement does not have to cope with problems like donor bleedthrough or directly excited acceptor fluorescence. This relaxes the requirements to the absorption and emission spectra of the donors and acceptors used. Moreover, FLIM-FRET measurements are able to distinguish interacting and noninteracting fractions of the donor, and thus obtain independent information about distances and interacting and noninteracting protein fractions. This is information not accessible by steady-state FRET techniques. Autofluorescence FLIM exploits changes in the decay parameters of endogenous fluorophores with the metabolic state of the cells or the tissue. By resolving changes in the binding, conformation, and composition of biologically relevant compounds FLIM delivers information not accessible by steady-state fluorescence techniques. 2ff7e9595c
Comments