Biosensor platform and method for simultaneous, multiplexed, ultrasensitive and high-throughput optical detection of biomarkers
By using broadband continuous spectral illumination source and micron-scale dense array spectrophotometer on biosensors, combined with an automatic focusing system and translation device, spatial and spectral resolution of nanoparticles is achieved, and the low signal-to-noise ratio and specificity in the prior art are solved. The poor performance problem realizes multiple detection and high-throughput diagnosis.
Patent Information
- Application Number
- CN202010260677.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Priority Date
- 2019-04-03
- Filing Date
- 2020-04-03
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2040-04-03
AI Technical Summary
The existing biosensing detection platforms are difficult to distinguish different types of nanoparticles, their aggregation states, and eliminate non-specific signals caused by substrates and impurities. Mechanical transduction cannot provide information on individual nanoparticles, resulting in low detection signal-to-noise ratio. Poor specificity and detection time are correlated with the number of markers.
The broadband continuous spectral illumination source and micron-scale dense array spectrophotometer are used, combined with an automatic focusing system and translation device to achieve spatial and spectral resolution of the biosensor surface, capture and analyze scattered signals through optical detectors, identify and count nanoparticles. In turn, biomarker concentrations are quantified.
The signal-to-noise ratio and determination specificity of biosensors are significantly improved, multiple detection and high-throughput diagnosis are realized, and the measurement time does not depend on the number of markers. It can detect multiple biomarkers simultaneously and quickly quantify biomarker concentrations.
Smart Images

Figure CN111795950B_ABST
Abstract
Description
Background of the Invention Field of the Invention
[0001] The present invention relates to a multiplexed biosensor for diagnostic applications that detects and quantifies physical changes that a biorecognition layer bound to a solid sensor undergoes when it interacts with a sample containing target molecules.
[0002] Description of Related Technology
[0003] Biosensors utilize the ability of some biomolecules (receptors) to specifically bind to (recognize) complementary biomolecules (ligands). The most typical interactions are complementary nucleic acid hybridization and antibody / antigen binding. Biosensors are popular in basic biological research, health science research, drug discovery and clinical diagnosis. Depending on the physical changes measured, biosensors can be classified as electrochemical biosensors, mechanical biosensors and optical biosensors. In EP3153844, a novel biosensing detection platform (biosensing platform) for ultrasensitive detection of protein biomarkers deposited on a surface is proposed. The key features of this biosensing detection platform are:
[0004] 1) Label each biomarker with a specific type of plasmonic nanoparticles (nanospheres, nanorods, nanocubes, nanoprisms, etc.).
[0005] 2) Plasmonic nanoparticles are deposited onto a functionalized suspended multi-dielectric substrate, which simultaneously allows to greatly enhance the weak scattered signal from the plasmonic nanoparticles (optoplasmonic effect) and weigh the mass of the nanoparticles.
[0006] The biosensing detection platform presents a dual transduction mechanism based on optical detection, where the plasmonic nanoparticles are optically detected using a standard dark-field microscope, and on mechanical detection, where the mass of the plasmonic nanoparticles is detected by measuring the change in the resonant frequency of the suspended mechanical substrate after the deposition of the nanoparticles.
[0007] This biosensing platform is very robust and has been tested with infectious, inflammatory and tumor protein biomarkers and it exhibits ultra-high detection sensitivity, up to six orders of magnitude better than the standard in routine clinical practice. However, the main drawback of this known biosensing platform is that it is not possible to:
[0008] i. Distinguish between different types of nanoparticles, since the light signal is averaged over all surface areas,
[0009] ii. distinguishing different states of aggregation of nanoparticles (monomers, dimers, trimers, etc.),
[0010] iii. Eliminate non-specific signals caused by light scattered by the substrate and other potential impurities,
[0011] iv. Extracting the basic spectral characteristics of the nanoparticles on the surface, since optical identification is performed using a standard dark-field microscope.
[0012] Furthermore, mechanotransduction cannot yield any information about individual nanoparticles, as only the overall mechanical properties of the sensor can be measured using this platform. SUMMARY OF THE INVENTION
[0013] The present invention provides a biosensing detection platform and method that solves the previously reviewed technical problems by simultaneously obtaining spatially and spectrally resolved information about the surface of the biosensor. Compared with the previous state of the art, using the invention described herein, it is possible to: i) greatly increase the signal-to-noise ratio of the biosensor, ii) significantly improve the specificity of the assay, iii) perform simultaneous detection of multiple biomarkers (multiplex detection), and iv) achieve high throughput suitable for diagnostic purposes.
[0014] Importantly, the time required for measurement and analysis does not depend on the number of biomarkers used in the multiplex.
[0015] The present invention comprises a biosensing detection platform provided with a biosensor, a broadband continuous spectrum illumination source, an optical detection system for detecting light scattered by the biosensor, and processing means for the detected light signal, wherein the optical detector is provided with a micrometrically dense array of spectrophotometers capable of simultaneously performing spatial and spectral analysis of the scattered light, and an autofocus system adapted to focus the scattered signal from the surface of the biosensor onto a camera. The platform is also provided with translation means for the optical detection system and / or the biosensor, such that the optical detection system and the biosensor can be moved relative to each other in three dimensions, and wherein the processing means is adapted to capture spatially and spectrally resolved light signals from a sample.
[0016] Spatial and spectral information from the captured light signal is processed in order to derive the position of the scattered particles on the surface of the biosensor and characterize them in terms of their spectral emission. Further processing is performed to classify each particle, in particular, i) determine whether the particle is one of the nanoparticles used as a label on the biosensor, ii) decide which label the particle belongs to when several labels are used simultaneously (multiple), iii) determine the state of agglomeration of the particle (monomer, dimer, etc.), and iv) discard all other particles that are not identified as labels. For each of the labels used on the biosensor, the number of particles identified as such a label is counted. Based on these numbers, the concentration of the biomarker in each sample on the biosensor is quantified and presented to the user of the platform as an output result. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] To complete the description and provide a better understanding of the invention, a set of drawings are provided. Said drawings show preferred embodiments of the invention and should not be interpreted as limiting the scope of the invention but merely as examples of how the invention may be implemented.
[0018] Figure 1 is a schematic diagram of a system according to the present invention.
[0019] Figure 2 Two schematic diagrams corresponding to embodiments with separate optical and illumination systems or coupled optical and illumination systems are shown.
[0020] Figure 3 is a flow chart of the method of the present invention.
[0021] Figure 4 Three different spectral fingerprints of three different types of nanoparticles are shown.
[0022] Figure 5 is a two-dimensional histogram of the values of two parameters: the normalized brightness of the particles along the horizontal axis, and the relative contribution of the "red" channel along the vertical axis.
[0023] Figure 6 The results for one particle class are displayed as a "heat map," ie, the particle numbers are gray-coded and represented in two dimensions to resemble the location on a biosensor.
[0024] Description of the Preferred Embodiments
[0025] refer to Figure 1 The biosensor detection platform of the present invention comprises:
[0026] Devices for illuminating biosensors with broadband continuous spectrum;
[0027] A device that simultaneously captures spatially and spectrally resolved light signals from a sample by using an optical detector with both spatial and spectral resolution;
[0028] A device to focus light from the biosensor surface onto the optical sensor;
[0029] - Means to move the biosensor and / or the optical head in three spatial coordinates to modify their relative positions.
[0030] The system comprises an optical system (OS) capable of capturing scattered signals from the sample, an illumination system (IS) for illuminating the sample at a glazing angle, an optical detector (OD) for simultaneous spatial and spectral analysis of the sample surface, an autofocus unit (AF) for fast and automatic focusing of the sample surface, and a motorization system (MS) capable of generating relative movement between the sample and the optical scanner.
[0031] Lighting System (IS)
[0032] The illumination system (IS) consists of, for example, a broadband VI-NIR light source coupled to an illuminator device. The term "broad" means that the light source exhibits a very broad spectral emission in the visible light (400nm-700nm) and / or in the near infrared spectral range (from 700nm to 1000nm). For the purposes of the present invention, the broadband light source can be a tungsten halogen lamp, a xenon lamp, a supercontinuum laser or a combination of multiple LEDs or lasers. The illuminator device collects the light from the broadband light source and illuminates the sample surface with a well-collimated light spot at an illumination angle.
[0033] Optical System (OS)
[0034] The main purpose of the optical system is to recover the weak scattered light from the sample surface in a very effective way. The optical system (OS) in the preferred design of the present invention can be an infinity-corrected optical system (infinity-corrected optical system) obtained by combining an optical objective lens with a tube lens. Although not optimal, the optical system (OS) can be designed to have only one optical objective lens (finite-corrected optical system).
[0035] As in Figure 2As shown in the schematic depiction of , the optical system (OS) and the illumination system (IS) can present separate or coupled light paths. Depending on the selected geometric configuration (separate or coupled), the illumination system can have different geometric configurations. When the illumination system (IS) is coupled with the optical system (OS), the illumination device can achieve Or Nelsonian illuminator. In contrast, when the IS and OS have separate optical paths, the illumination system can use a lighting angle illuminator.
[0036] In the same way, the optical system presents significant differences depending on the geometrical configuration of the OS and IS relative to each other (coupled or separated). When the optical system (OS) is coupled to the illumination system (IS), some elements that enable the optical path of the illumination to be separated from the optical path of the light scattered by the sample must be implemented in the optical system (OS). This condition is usually met by installing optical objectives and beam splitters in the optical system (OS) that are specially designed for darkfield microscopy applications. On the contrary, when the optical system (OS) is separated from the illumination system (IS), the optical system can be designed with standard optical elements used in brightfield applications.
[0037] Optical Detector (OD)
[0038] The optical detector captures the scattered light collected by the optical system, and it has the ability to simultaneously resolve the light signal from the sample surface with high spatial and spectral resolution. In order to perform spatial and spectral analysis of the sample simultaneously, the optical detector must be a micron-scale dense array of spectrometers. The dense array gives the ability to analyze different spatial positions of the sample simultaneously, and each spectrometer of the array allows spectral analysis in each point of the sample. The optical detector described here allows the sample to be analyzed simultaneously in spatial and spectral coordinates. The dense array of spectrometers can be implemented as an array of photodetectors coupled to an array of filters arranged in a mosaic pattern, ii) multiple arrays of photodetectors coupled to dichroic filters, or iii) a vertically stacked array of photodetectors capable of detecting different spectral bands at the same spatial position. These types of optical detectors are viable technical solutions for simultaneous spatial and spectral analysis of samples with sub-micrometrical resolution and in multiple spectral bands (usually at least 2 and not more than 30). Preferentially, optical detectors with a high dynamic range should be used, allowing to image nanoparticles associated with different biomarkers with only one capture, so that the time required for the measurement process is not increased in case more than one biomarker is used.
[0039] Auto Focus System (AF)
[0040] In order to focus on the sample surface in a fast and automatic way, the optical system is equipped with an autofocus system (AF); the automatic focusing of the sample can be achieved via hardware or via a software solution. Hardware-based autofocus systems, which are generally referred to as active autofocus systems in the literature, are usually obtained by projecting a laser onto the sample (the laser is part of the AF) and collecting the laser reflected from the sample with a differential photodetector. By measuring the position of the reflected light on the photodetector, the distance of the sample surface from the best focus position can be quantified; therefore, the best position is restored by moving the biosensor, the optical system or a part of it or both hardware components, using one or more motorized stages that can modify the relative distance between the sample and the head of the optical scanner. In order to accurately focus on the sample surface, the typical resolution of the movement needs to be much lower than the depth of field of the high-resolution optical objective (which can typically reach 200nm for a 100x optical objective). Many different technical solutions can be implemented, such as stepper motors or DC motors or piezoelectric actuators.
[0041] The software-based autofocus system captures several images of the sample at different relative distances between the sample and the optical head; the capture of the images can be performed with the same optical detector used for the optical analysis of the sample or with a dedicated optical detector. The different captured images are analyzed with a dedicated algorithm that calculates the contrast of each image and can quantify the distance of the sample surface from the best focus position. The best position is finally restored with one or more motorized stages that are able to change the relative distance between the sample and the optical head of the scanner.
[0042] Translation device (MS)
[0043] Because the measurement needs to be done over a very wide surface area (several thousand cm 2 ), so the system also requires the use of a motorized system that can modify the relative position between the biosensor surface and the optical system.
[0044] There are many different experimental configurations that can be used to accomplish this task:
[0045] A. The sample moves along two axes; the optical system remains stationary.
[0046] B. The optical system moves along two axes; the sample holder remains stationary.
[0047] C. Both the optical system and the sample are moved, for example, the optical system along one axis and the sample holder along a second axis.
[0048] Biosensor (BS)
[0049] The biosensor comprises a functionalized substrate and functionalized nanoparticles. The target analyte is an element to be detected from a sample, in particular a biological sample. The target analyte can be of any nature, such as an organic or inorganic molecule (drug, hormone, cholesterol, etc.), a biomolecule (peptide or protein, nucleic acid molecule, growth factor, biomarker, etc.), a cell (protozoan cell, bacterial cell, fungal cell, eukaryotic cell) or a cell fragment (cell fragment) (bacterial wall, organelle (such as mitochondria, cell vesicle, etc.)) or a virus.
[0050] The recognition element that makes the substrate surface functionalized can be any element that can recognize and specifically bind to the target analyte. In this sense, the recognition element can be an antibody (polyclonal or monoclonal antibody), a receptor (cell surface receptor such as an opioid receptor), a peptide (e.g., an opioid peptide), a protein (e.g., a lectin (lectins)), a carbohydrate (e.g., a lipopolysaccharide O-antigen), a nucleic acid (DNA or RNA sequence), a cell (protozoan cell, bacterial cell, fungal cell, eukaryotic cell), a microorganism or a part thereof (e.g., a bacterial wall, an organelle (e.g., mitochondria, cell vesicles, etc.)). In a preferred embodiment of the present invention, the recognition element is an antibody, more preferably a monoclonal antibody.
[0051] The nanoparticles must naturally have plasmonic properties. In principle, any type of nanoparticles having plasmonic properties can be used. Thus, the nanoparticles can be, for example, gold nanoparticles, silver nanoparticles or nanoparticles of plasmonic metamaterials, such as, but not limited to, titanium nitride and non-stoichiometric oxides (e.g., non-stoichiometric vanadium, titanium and aluminum oxides).
[0052] In addition, the nanoparticles can take a variety of forms or structures, such as nanospheres, nanorods, pointed nanorods, nanoshells, nanocages / frames, hollow nanospheres, tetrahedrons, octahedrons, cubes, icosahedrons, rhombic dodecahedrons, concave nanocubes, tetrahexahedra, obtuse triangular bipyramids, trisohectahedra, and nanoprisms.
[0053] The nanoparticle comprises at least one detection element bound thereto, which can in particular bind to the target analyte.The detection element can be any type of element that can bind to the target analyte, so in principle its properties can be the same or similar to those of the recognition element.
[0054] The function of the detection element is to detect the presence of a target analyte captured by the recognition element fixed on the substrate surface. Therefore, if the target analyte is present in the sample being analyzed, the nanoparticles will only bind to the substrate with the aid of the detection element bound to it. In this case, the recognition element can bind to the target analyte, which is then detected by the detection element in the sandwich arrangement. The absence of the target analyte in the sample causes the recognition element not to bind to the target analyte, and therefore detection by the detection element will not occur.
[0055] In summary, if the target analyte is present in the sample, even at ultra-low concentrations, it can be detected and quantified based on the scattering spectrum produced by each individual nanoparticle. If the target analyte is not present in the sample, there will be no detectable plasmon effect on the substrate because the nanoparticles will not be present.
[0056] Analysis of spatially and spectrally resolved information
[0057] The optical system and detector together with the translation device are used to obtain spatially and spectrally resolved information about the biosensor. This information is referred to as "images" below, keeping in mind that these "images" can contain more and / or different spectral information than "typical" microscope images. Analysis of these images provides the concentration of biomarkers detected in each sample on the biosensor. Three aspects are used to minimize the time between image acquisition and results:
[0058] 1. The main part of the analysis runs in parallel with the image acquisition.
[0059] 2. Analyze several images in parallel, since the analysis method allows to process the images independently for most analyses.
[0060] 3. Implement highly effective image analysis methods.
[0061] The core of the analysis lies in the identification, classification and counting of particles (see Figure 3). Each image is read and, if necessary, corrections (background, inhomogeneities, etc.) are applied. The particles are first located in each image analyzed, they are characterized (brightness, emission spectrum, etc.), the results are used to classify them (nanoparticle monomers, clusters, dust, etc.), and finally they are counted by category. These numbers constitute the main results for each image; in addition, additional results are derived that allow the quality of the measurement to be controlled (correct focus, etc.). From the number of different particles in each image, the concentration of the corresponding biomarker in the corresponding sample is derived.
[0062] Preferred Embodiments
[0063] In a first implementation, a white light source based on LED is used for illumination. The light source is selected so as to cover the scattering spectrum of the particles of interest, generally from blue (450nm) to red (700nm) for spherical gold nanoparticles. Because smaller particles scatter less light and their spectrum shifts toward blue, preferably, a light source with a stronger emission in the blue-green region is selected to (partially) compensate for this effect. This compensation facilitates imaging of all particles with only one setting of an optical detector. In order to achieve high optical resolution and effective light collection of the optical system, a dark field microscope objective with a high numerical aperture (NA) is used, such as a 50x / 0.8 objective. In order to ensure accurate characterization and classification of particles on the biosensor over the entire field of view (FOV), an objective that is well corrected spherically and chromatically, such as a plan apochromat, is used. An infinity-corrected objective is used to ensure flexibility in the combination with other components of the optical system.
[0064] In a first implementation, a CCD sensor with RGB filters is used; in a second implementation, a CMOS sensor with RGB filters is used; the preferred embodiment utilizes a CMOS sensor with 16 different spectral bands arranged in a mosaic pattern. Alternatively, multiple monochrome CCD / CMOS sensors coupled with dichroic filters can be implemented as optical detectors, or even sensors based on Foveon technology.
[0065] From the above, an obvious alternative to the use of a color camera or a sensor with multiple spectral bands is a combination of a monochrome camera with external filters (with a set of fixed filters; with a single but spectrally tunable filter, etc.) or with variable illumination (switching light source; light source plus monochromator, etc.). In all of these alternatives, the camera will take one image for each filter or illumination setting, resulting in an equivalent image stack. However, due to the necessity of taking more than one image at each scanning position and changing the filter or illumination setting between images, these alternatives are significantly slower, resulting in very low throughput that is not suitable for diagnostic purposes. An advantage of the method of the present invention is that the required spatial and spectral information is obtained in parallel rather than continuously to ensure the highest possible data acquisition speed. Importantly, the optical detector used in the present invention needs to perform spatial and spectral analysis of the sample surface simultaneously. The platform of the present invention allows the measurement of approximately 100 samples in a few minutes, which is at least 10-100 times better than traditional microspectrophotometric techniques.
[0066] The geometric dimensions of the sensor are selected to approximately match the size of the image transmitted by the optical system (i.e., microscope objective). In standard microscopes, this size is called "field number" (FN), and is equal to the aperture diameter of the eyepiece (eyepieces) in millimeters. If the sensor is too small, it cannot make full use of the available field of view given by the objective lens, and if the sensor is too large, it is only partially used. Because the aberration of the optical system tends to increase toward the image boundary, the sensor size slightly smaller than the FN is usually the best choice to ensure high image quality and thus reliable detection, characterization and classification of particles. In the current implementation, the sensor diagonal is about 18mm, and the FN is 22mm. Because the different types of nanoparticles used in the multiple channels can usually have very different scattering efficiencies and will therefore significantly change (e.g., 30 times or more) in brightness, a sensor with a large dynamic range is used to allow all particles to be imaged with a single camera setting. A large dynamic range usually requires a sensor with a sufficiently large image pixel and a low noise level.
[0067] The combination of microscope objective and optical detector is selected so as to allow imaging of particles smaller than 1 μm (e.g., having a diameter in the range between 50 nm and 150 nm) with an optical resolution better than 1 μm (e.g., typically between 0.5 μm and 0.9 μm) and resulting in sufficient pixels per particle in the image (e.g., 10 pixels per μm, referring to the object plane). Thus, the optical detector typically has between 4 and 12 megapixels and a diameter of approximately 10×15 mm. 2 of size (e.g. of the so-called 1"-sensor type).
[0068] In the present invention, the number of spectral bands used has little effect on image acquisition or on data analysis; what is important is that the acquired image can be arranged so as to have spatial information in two dimensions (image pixels) and spectral information in three dimensions (image layers). In the example (see Figure 4 ), assuming that a biosensor containing three types of nanoparticles with three different spectral fingerprints is imaged. The image sensor delivers a stack of five image layers, each corresponding to a spectral range. Using these five layers, a simplified version of the scattering spectrum shown on the right will be obtained at each image pixel (only five data points). The more spectral layers, the better the reproduction of the scattering spectrum and the easier it is to distinguish between the spectra of different particles.
[0069] In the following, an RGB camera is used as an example to explain the image analysis method used in the present invention; however, it should be understood that even if we talk about "color" or "color component" in the following, "color" always refers to the complete spectral information from all layers available at each image pixel, and "color component" refers to the data from one of these layers. In the same way, the combination of spatial and spectral information is called an "image", regardless of the number of spectral layers.
[0070] The parameters of the camera are adjusted to produce images with good exposure with respect to black and white levels, white balance and color reproduction. The settings are chosen so that the nanoparticles used as biomarker labels can be imaged well, with a signal above the noise level and below the saturation region (by adjusting sensitivity and exposure time) and resulting in good differentiation between different particles in terms of scattering spectra ("color") (by adjusting the color correction matrix).
[0071] Typically, the biosensor to be scanned includes different areas corresponding to different samples, much like a multi-well plate, with one well per patient sample.
[0072] Spatial resolution on the biosensor is achieved by two means: the optical detector (i.e., camera sensor) itself provides spatial resolution, and the biosensor and optical system are moved relative to each other. In the current implementation, the biosensor is moved relative to the stationary optical system by means of a two-axis motorized stage.
[0073] Typically, more than one image is taken for each area corresponding to the same sample, however, the images taken typically do not completely cover the sample area. In a general scan, the number of images taken per sample and their positions within the sample area may be the same for all sample areas of the biosensor.
[0074] However, this is not necessarily the best option; the number and position can also be selected individually for each sample, for example in order to take more images for samples with low concentrations of biomarkers to improve the statistical robustness of the measurements. The total number of images taken by the optical platform during a scan of the biosensor can range from one to several thousand. The subdivision of data acquisition in separate images has an important advantage: the analysis of these images (see the corresponding section) can be performed in parallel with the scanning, i.e., while the optical platform continues to acquire images. This allows a higher sample throughput of the platform.
[0075] The images captured during the scan are saved in a common image format (e.g., TIFF or JPEG), typically RGB images with 8 bits per color channel. In most cases, JPEG is preferred because the resulting files are smaller and can be written and read faster. On the other hand, JPEG uses lossy compression, which particularly affects color representation. Because the spectral characterization of the particles on the biosensor is an important aspect of the present invention, only a fairly mild JPEG compression (i.e., a high "quality factor") is used to minimize potential distortions in the color representation. Optionally:
[0076] • The image can be saved with a larger color depth, e.g. as a 16-bit TIFF image, to avoid artifacts in the image in case of particles with lower scattering intensity.
[0077] • Images can be saved as camera raw data. This preserves the full dynamic range of the camera sensor (typically 12-14 bits) and the linear correlation between the amount of scattered light and the signal measured by the sensor.
[0078] In the case of cameras with multiple spectral bands, the manufacturer's proprietary image format usually has to be used; in addition to the RAW format, it is also possible to base TIFF as a container.
[0079] In order to reduce the time and amount of memory required to store images and the time required to analyze them, the captured images are binned. Typically, 2x2 binning is applied, i.e., four sensor pixels produce one image pixel. A quadratic interpolation between the pixels is used to compute the binned images. This approach achieves better differentiation between particles in terms of color compared to the alternative of directly using a sensor with fewer pixels without subsequent binning of the images. Therefore, in the current implementation, image binning is applied when using a camera with 12MP; images from a 4MP camera are not binned because the remaining spatial resolution is too low.
[0080] Typically, the images are first stored locally on the hard disk or similar storage device of the computer. Alternatively, the images may be transferred directly to a different computer or storage system within the network of computers of the platform. Furthermore, if instead the analysis of each captured image (see the following paragraph) is performed directly while the image is still in the memory (RAM) of the computer, then storing the images to a storage device (e.g., hard disk) may be omitted.
[0081] For the analysis of the images from the scans, an analysis program consisting of scripts compatible with Matlab (available on the market from MathWorks) and GNU Octave (free software) can be used. In addition, some or all of the analysis can be implemented in hardware, for example, using an FPGA (field programmable gate array) directly connected to the optical detector. This can reduce the required analysis time. Before the actual analysis, the parameters of the data acquisition ("scan") required for the analysis are obtained; usually, they are read from a file or database, or manually entered by the user. These parameters should include the number and position of the sample and the individual images on the biosensor, the type of biomarkers present in the sample, the type of nanoparticles and biosensor substrates used, information about the sample volume, lighting, and camera and image capture settings, etc. In the current implementation, this information is automatically saved in two text files during the scan; one file contains information about the settings of the camera, lighting, and autofocus, and the second file contains information about the geometry of the sample on the biosensor, the geometry of the image in the area of each sample, and the sample that is actually scanned (i.e., all samples or their subsets). In addition, the correspondence between the sample and patient information on the biosensor must be stored (e.g., which well / sample on the biosensor corresponds to which patient ID); this information is irrelevant to the analysis itself, but is of course essential for the diagnostic use of its results. In a clinical setting, this information is usually stored in a central hospital information system (HIS). In the current embodiment, the information is compiled by the person who prepared the biosensor and stored in a folder on a network drive, from which it can be read by both the platform of the present invention and the analysis software.
[0082] Analyze all images (or a subset of all images) acquired with the optical platform. On the computer, the analysis can be performed strictly sequentially (one image at a time), or several images can be analyzed in parallel (using multiple threads of the computer, for example, one image per thread). Usually, the number of parallel threads (= the number of cores or the number of logical processors) close to the maximum value available on the computer is selected to reduce the total time of the analysis. In the current implementation, the analysis is run on the same computer that controls the data acquisition of the platform. The possibility of obtaining information of interest on the biosensor by analyzing different images independently is a major advantage of the present technology: in this way, the most time-consuming task (analysis of each image) can be easily parallelized, and the parallelization can be increased in a direct, effective and economically feasible way (computers with more cores, several CPUs, several computers in the network, etc.). Optionally, if the images are already stored in the network, the analysis can be run on a computer different from the computer that controls the platform. Similarly, the analysis of all images can be divided between several computers in the network, so that each computer analyzes a subset of the images, or both storage and analysis can be done using cloud services. Parallelization can be per image (one image per thread) or by subdividing each image and analyzing sub-images in parallel. Analysis can be performed after the entire biosensor is scanned, or it can be run in parallel with the scan (e.g., whenever all images of a certain sample are taken) so that results are obtained as quickly as possible after the scan.
[0083] During analysis, the following steps are performed for each image: First, the image data is read from the storage device (or directly from the computer's memory). Then, if the corresponding option has been selected by the user, or activated by default, or automatically activated based on information about the scan (type of camera sensor used, imaging parameters, etc.), the image correction and / or transformation is then calculated.
[0084] Background correction of the image is performed to adjust the black level. In a typical image from the platform of the present invention, the particles of interest appear as individual small bright spots on a dark background. Depending on the substrate, the type of camera sensor and the acquisition parameters, the background may not appear "black" (i.e., the sensor value is close to zero, the "ideal" case), but "gray" (with or without a certain tint). Estimation of the sensor value (e.g., based on a histogram of all pixels) allows correction for signal offsets and potential tints. This simplifies the characterization of the particles in the later steps of the analysis.
[0085] The image is then corrected for potential non-uniformities in brightness. Such non-uniformities may be due to illumination, light collection, camera sensor, etc. The general effect is that the center of the image appears brighter than the borders. Correction for these effects may be necessary to ensure correct characterization and classification of the particles in later steps of the analysis.
[0086] Modification of the gamma curve of the image is performed to adjust the dynamic range. The "gamma curve" refers to the dependence of the pixel values in the image on the actual (physical) amount of light detected. In standard images (JPEG, TIFF, etc.), this dependence is non-linear, so that with increasing light, the pixel values increase more slowly than proportionally. In the analysis, the inverse function of the "gamma curve" is used to correct for this non-linear dependence, so that the obtained pixel values are again proportional to the amount of scattered light.
[0087] The image data is smoothed to reduce image noise. As stated above, the actual optical resolution is usually coarser than the pixel pitch on the camera sensor. This means that image noise (from the camera sensor, due to low light intensity, etc.) can be reduced by smoothing without affecting the resolution. Reduction in image noise again improves the correct characterization and classification of particles.
[0088] In the case of a standard color camera, a transformation to a different color space can be performed, such as from RGB to HSV, L*a*b, etc., which can simplify the interpretation of the results from the particle characterization and lead to a more precise distinction of particles in terms of color. For example, the L*a*b color space contains one channel for light intensity (L), while the other two axes describe color; in comparison, RGB's channels are a mixture of color and brightness.
[0089] If the color space used does not contain a dedicated brightness value (e.g., in RGB), such a value is calculated by summing the color channels of each image pixel. The resulting grayscale image is referred to as the "brightness image" in the following. This brightness image is usually normalized, with "0" and "1" corresponding to the lowest and highest brightness levels in a given image format (8-bit RGB: (0|0|0)→0, (255|255|255)→1).
[0090] If the color space used does not contain relative color values, then such relative color contribution can be calculated for each pixel by dividing each color channel by the sum of the channels, e.g. the relative contribution of the "red" channel is R rel =R / (R+G+B).
[0091] In case of using a camera with multiple spectral bands, the scattered intensity measured in each wavelength range of the sensor is used instead of the RGB values. In both cases, the analysis method does not change significantly; for example, the three values of the RGB image are simply exchanged for an array of elements representing different wavelength regions. The number of elements is usually greater than three to obtain spectral resolution. However, the combination of just two well-selected wavelengths can also be a suitable choice for precise discrimination between different particles.
[0092] Potential particles are then located in the image: a grayscale test pattern representing the general shape of the nanoparticles in the image from the optical platform is calculated, taking into account the actual image acquisition parameters used (optical and sensor resolution, magnification, etc.). The shape can be represented by, for example, a two-dimensional Gaussian function, an Airy pattern, a simple disk shape, etc. In the current implementation, the sum of two Gaussian functions (G1+(-G2)) is used to approximately match the general "ring" shaped scattering pattern of individual gold nanoparticles. As a variation, a pattern with color (e.g., one pattern for each RGB component or for each channel of a camera with multiple spectral bands) can be used instead of a single grayscale test pattern. This can improve the correct identification of potential particles in the case where the shape depends on the wavelength (e.g., a Gaussian shape in the green spectral range and a ring shape in the orange spectral range).
[0093] Calculate the standardized two-dimensional cross-correlation between the brightness image and the test pattern. The cross-correlation image has a value close to the value of the image area similar in shape to the test pattern. Instead of using the brightness image, the cross-correlation between the test pattern and, for example, one of the color channels of the image (or a linear combination of several channels or some other function) can also be calculated. If, for example, a certain pattern is more obvious in one channel than in other channels, this is a suitable method. Calculate a binary mask for the image. For pixels with a cross-correlation value above a certain threshold (e.g., >0.6) and a relative brightness within a certain range (e.g., >0.03 and <0.98, i.e., above noise and below saturation), the mask is equal to 1 (=true). All other pixels are zero (=false). Instead of using a binary mask ("0" or "1"), a mask with continuous values can be used, such as a grayscale mask with values close to 1 (if the constraints are well matched), and a grayscale mask with values close to 0 (if the constraints are not matched), and all values in between are possible. The given thresholds for correlation and brightness are only examples; based on actual measurement results, more appropriate values can be selected, or thresholds for different and / or additional parameters (color, signal-to-background ratio, etc.) can be selected. The binary mask is then multiplied pixel by pixel with the cross-correlation image. In the resulting grayscale image, local maxima (= pixels with a higher value than the values of all directly adjacent pixels) are located. Instead of using the product of the binary mask and the cross-correlation image, the positions of potential particles (the positions of areas with a value of "1") can be derived only from the binary mask. Instead of using the product, other functions are used, for example, taking the square of the cross-correlation image, etc. The positions of these local maxima are considered potential particle positions. A variation is that no cross-correlation with the test pattern is used at all, but a particle search based on thresholding techniques (for example, everything above a certain background value is considered as a potential particle), image-based Hough transform (for example, finding a circular object of a certain size), etc. is used.
[0094] An important aspect of the present invention is that each particle is individually characterized; this allows for the classification of particles in the following analysis steps.
[0095] The characterization was performed as follows:
[0096] a) The parameters of greatest interest are the brightness and the scattering spectrum ("color") of each particle. To obtain average values of these parameters (i.e., averaged over the area of the particle), the corresponding images (brightness images, images with relative color contributions, etc.) are convolved with a suitable kernel, such as a 2D Gaussian with dimensions close to the expected particle size or with a disk-shaped pattern.
[0097] b) The filtered image evaluated at the previously calculated potential particle positions.
[0098] Variants: (a.-b.) In cases where the number of particles per image to be characterized is "low", it may be computationally inefficient to compute the convolution of the entire image with the kernel to extract the mean. Instead, it is more efficient to extract small regions around each particle directly from the image and use them to derive the mean. What is considered a "low" value depends on the imaging parameters, the computer's hardware, and the implementation of the analysis routine. For a given setup, the crossover point can be determined so that one or the other approach can be used automatically.
[0099] c) Additional properties of interest are the size and shape of the particle, such as the FWHM in the brightness image, the degree of correlation with the test pattern and the local shape at the center of the particle. The latter parameter allows to distinguish between a donut shape (= indentation at the center) and a Gaussian (or similar) shape (= maximum at the center). The analysis used in the current implementation uses a value proportional to the discrete Laplace at the center of the particle; this choice allows an efficient implementation based on a simple convolution of the image with a 3x3 kernel.
[0100] d) Other additional properties of interest are local particle density (e.g. based on nearest neighbor distance) and further spectral properties (e.g. differences or ratios of spectral components).
[0101] e) As a result, the characteristics of each potential particle are obtained (their positions are known from the previous steps). For each image, this intermediate result is represented as a table with one row per potential particle and as many columns as characteristics are derived, see Table 1 for an example of a diagram representing the results of the characterization step from one image. Each row of the table corresponds to a particle, x and y are its coordinates in the image, I is its brightness, and R, G, B are the relative color contributions. Depending on the analysis, more columns are added for additional characterization parameters.
[0102] Table 1
[0103] Particle classification
[0104] Particle sorting allows only particles that are most specific for the biomarker and essential for multiplexing to be considered in the analysis.
[0105] The work is classified as follows:
[0106] a) If desired, certain characteristics can be used to exclude particles from further steps of the analysis. For example, if only monomers of nanoparticles are of interest and the monomers show a donut shape as a distinguishing feature, this parameter can be used to eliminate all non-donut particles.
[0107] b) Based on the information from previous experiments, different classes of particles are defined based on their properties (brightness, "color", shape, size), for example, the different classes of particles are as follows:
[0108] (1) Noise (very low brightness),
[0109] (2) residues from biosensor manufacturing (certain spectral patterns),
[0110] (3) Dust (different spectral patterns),
[0111] (4) Individual nanoparticles (different spectral patterns, brightness, shapes, etc.),
[0112] (5) Nanoparticle clusters (dimers, trimers, etc., based on spectral patterns, brightness). For these clusters, their brightness often depends systematically on the brightness of the corresponding individual nanoparticles. This dependence is used to improve classification.
[0113] c) In case of multiplexing (=simultaneous detection of various biomarkers, each labeled with a different nanoparticle), additional classification groups are defined to account for all nanoparticles used.
[0114] d) The necessary classification parameters / rules (which combinations of characteristics correspond to each class) can be given in advance (based on previous measurements) or can be derived from the measurements to be analyzed themselves. In the second case, the analysis is usually based on the combined results of all images from the measurement, rather than on each image individually (to improve the statistics and ensure that the classification is consistent among all images).
[0115] e) Classification essentially consists in a segmentation of the parameter space. Its rules can be defined by the user (e.g. manual segmentation) or can be derived automatically (e.g. cluster search, k-means method, segmentation algorithm, machine learning algorithm, etc.). In the case of an RGB camera, an example of classification is that all particles with a normalized brightness in the range [>0.1, <0.7] and a relative contribution of the "red" color channel in the range [>0.0, <0.33] are considered to be residues from manufacturing.
[0116] In the current implementation, classification rules are derived as follows:
[0117] i. Select two parameters from the characterization step and compute a 2D histogram of these two parameters from all potential particles. In the case of an RGB camera, the normalized brightness and the relative contribution of the "red" channel are usually used as parameters.
[0118] ii. In such a two-dimensional histogram, particles of the same type (e.g. individual nanoparticles) will form a kind of "peak" (= "dense" area of the histogram or of the parameter space in general) because all of them have similar brightness and color values. Due to differences between individual particles and due to noise and / or lack of precision in the measurement, these particles are similar but not identical.
[0119] iii. In a typical measurement, a main "peak" corresponds to individual nanoparticles (monomers). Based on knowledge from previous measurements, the position of this "peak" in the histogram of the data to be analyzed can be automatically identified, searching for the local maximum of the histogram within a given parameter range, e.g., normalized brightness in [0.1, 0.3] and the relative contribution of the "red" channel in [0.35, 0.45].
[0120] iv. Once this position of the monomer is known, the position of the dimer, trimer, etc. can be estimated based on the knowledge that they are approximately twice, three times, etc. brighter than the monomer. Again, the estimates are used to guide the search for the actual local maximum.
[0121] v. The peaks for monomers, dimers, etc. tend to overlap slightly. To assign a particle to one peak or the other, a "boundary" between the two peaks is defined: the routine searches for a (straight) line that best defines the "valley" between the two peaks.
[0122] vi. The positions and dividing lines in the histogram between monomers, dimers, trimers and larger agglomerations of nanoparticles are identified as described above. The rule for classifying a particle as, for example, a dimer is then that the parameters of the particle correspond to points on the "right" (= higher normalized brightness) of the "border" separating monomers and dimers in the histogram and to the left (= lower brightness) of the "border" separating dimers and trimers. In addition, the data points must be within a certain range with respect to the contribution of the "red" channel (e.g., greater than 0.34 and less than 0.6).
[0123] vii. In the current implementation, nanoparticles are classified as monomers, dimers, trimers, and "larger agglomerates" (ie, N ≥ 4).
[0124] viii. In addition to the four rules for automatic adaptation for monomers, dimers, trimers and larger agglomerations of nanoparticles, fixed rules are defined for four more types of "particles":
[0125] (1) Noise: all “particles” with a normalized brightness below 0.1 that do not belong to the defined group of nanoparticles.
[0126] (2) Residues with low brightness: "red" is lower than 0.33, and the brightness is in the range [>0.1, <0.7].
[0127] (3) Residues with high brightness: "red" is lower than 0.33, and the brightness is in the range [>0.7, <1.0].
[0128] (4) Saturated particles: contain image pixels with a brightness value equal to 1 (= saturation).
[0129] f) Classification can be adapted manually or automatically in case of differences between nanoparticle batches, lighting variations, detection parameter changes, etc.
[0130] g) As a result, a classification of all potential particles is obtained. The representation of this result is based on the table generated in the characterization step, with the addition of a column with the classification results, see Table 2 for an illustration showing the result from the classification step for one image. Compared to Table 1, a column is added to the right, indicating the "class" to which the corresponding particle belongs. In this example, the class is represented by an integer.
[0131] Table 2
[0132] x y I R G B ... ... category 12 36 0.13 0.45 0.33 0.22 4 17 20 0.57 0.42 0.34 0.23 7 18 102 0.02 0.33 0.37 0.30 1 ...
[0133] Particle counting
[0134] The method described here is a "digital" technique: instead of integrating the light signal from the labeled biomarkers over a certain area of the biosensor, the actual number of detected labeled biomarkers is counted. This makes the method much more robust and independent of image acquisition parameters (such as brightness or spectrum of illumination) and independent of variations in biosensor substrates, nanoparticles, etc.
[0135] Particle counting works as follows:
[0136] a) The number of particles in each classification group is counted.
[0137] b) The "count" can be direct (one particle, one count) or weighted (e.g. weighted by the brightness of each particle in a group, or by an estimate of the probability of correct classification, etc.).
[0138] c) As a result, the number of particles per classification group is obtained for each image. The result can be represented as a table, for example with one row per image and one column per classification group.
[0139] In the case of multiplexed channels, it may happen that some groups overlap significantly in parameter space, e.g. clusters of less bright nanoparticles (dimers, trimers, etc.) with monomers of different (brighter) nanoparticles. In this case, the calculated numbers have to be corrected:
[0140] a) Measurements where only one type of nanoparticle is present are used to calculate the ratio between monomers, dimers, trimers, etc. for each type of nanoparticle used in the multiplex channel.
[0141] b) For these particles, calculate which fraction of them will fall into the region of the parameter space belonging to the other type of nanoparticles used.
[0142] c) Starting with, for example, the lowest-brightness monomers, their number is used to estimate the number of corresponding dimers, trimers, etc. and how many of them occur in regions of the parameter space corresponding to other nanoparticles.
[0143] d) These numbers are used to correct for the number of particles counted in a certain area, i.e. the number of clusters is subtracted from the uncorrected counts.
[0144] e) Repeating the process for the remaining nanoparticles, for example in order of increasing brightness.
[0145] Deformation: In case of considerable overlap between dense areas of particles from two (or more) types in the histogram, the assignment of a simple boundary between them may lead to a large number of misclassified particles. This can be reduced with an alternative approach: an appropriate sum of functions is used to fit the histograms or density distributions obtained from all images, with at least one component for each particle type of interest (a "peak" in the histogram). Then, for all particles per image (or all particles corresponding to the same sample), the weights of these components that best match the histogram of that image (or sample) are determined, and these weights are used instead of the previously interpreted particle counts.
[0146] Quality Control
[0147] To assess the overall measurement quality and the quality of each image in particular,
[0148] Calculate the areas of the image where saturation occurs (i.e., areas where the normalized brightness is close to 1),
[0149] • Derive a value representing the degree of focus of the image (eg, the average measured size of individual nanoparticles).
[0150] In addition to general quality control, these values can be used to guide the calculation of analytical results, see the following steps.
[0151] Overall analysis results
[0152] In case more than one image is acquired for the same sample, a suitable statistical value is calculated from all images of the same sample (e.g., average or sum of values from the images, trimmed mean, median, quantile, etc.). In the current implementation, the trimmed mean with a symmetric trimming of 40% is calculated from all images belonging to the same sample.
[0153] In the calculation of this value, parameters related to the quality of the image can be used to guide the selection of the most representative images of the sample, such as omitting those images that have saturation occurring in too large an area (e.g., saturated area > 10%) or that are not well focused (e.g., FWHM of monomers > 1 μm).
[0154] For each sample, a value is calculated that relates to the number of biomarkers present in the sample. In the case of multiple channels, one such value is calculated for each biomarker. A suitable choice for this value is the number of individual nanoparticles used as labels. The results can be presented as the average number of particles per image, the particle density (e.g., per mm 2 The latter representation is preferred because it is the most straightforward to interpret, meaning that in a certain (known) number of patient samples used on the biosensor, that number of biomarkers is detected.
[0155] The values presented to the user are provided with an appropriate measure of their uncertainty, for example based on their standard deviation or their coefficient of variation (CV). These uncertainties can be estimated from the variation among different images from the same sample and / or from the variation within a single image. They can be reported directly as a number (N±ΔN particles) and / or used to indicate whether the result is reliable.
[0156] If the uncertainty of the results for a sample is above a certain limit, the analysis software can feed this finding back to the biosensor detection platform. In this case, the scanner can acquire additional images of the corresponding sample to verify whether consistent results can be obtained. As another step, the analysis results can directly guide the entire scan, that is, for each sample, images are acquired until certain quality parameters are met or an upper limit on the time or number of images per sample is reached.
[0157] The overall result for each sample and each biomarker can be qualitative (presence or absence of the biomarker) or quantitative (concentration of the biomarker).
[0158] Proof-of-concept experiments, multiplex spatial and spectral analysis of plasmonic nanoparticles for biosensing applications Analysis:
[0159] Preparation of a biosensor for detecting two biomarkers, duplex interleukin IL-6 and IL-10 A biosensor with 96 holes was used, and eight different biomarker concentrations (from 1fg / ml to 1ng / ml, plus negative control) were repeated 12 times each. A Si-based multi-dielectric substrate with a size of 120mm x 80mm was used for the biosensor. After silanization of the surface, a self-assembled monolayer of a 1:1 mixture of two capture antibodies based on the biomarker of interest was grown. The division of the biosensor in 96 rectangular holes was achieved with a removable upper structure (GraceBio). Spherical gold nanoparticles (GNPs) (NanoPartz) with diameters of 100nm and 80nm were functionalized with IL-6 and IL-10 detection antibodies, respectively. A 1:1 mixture of two types of functionalized GNPs was prepared. For the sample, a buffer solution (PBST-25% FBS) was added with a serially diluted biomarker (1:10) to obtain a final concentration from 1ng / ml to 1fg / ml, plus a negative control. 200μl of solution was used per well. The distribution of the samples is shown in Table 3; concentrations are in g / ml, "0" is the negative control. Each concentration (rows 1...8) was repeated 12 times (columns 1...12).
[0160] Table 3
[0161]
[0162] After two incubation steps (first with sample and then with GNPs), the 96-well superstructure was removed. The biosensor substrate was washed several times and finally blown dry with dry nitrogen.
[0163] In this experiment, a platform according to the invention was used having the following components:
[0164] Dark field microscope objective lens 50× / 0.8
[0165] Dark field EPI illumination using high power white LED light source
[0166] Camera with 12-megapixel CMOS RGB sensor
[0167] 2×2 binning of the image applied before storing in JPEG file format
[0168] In each well of the biosensor, 13×13 images were taken. The acquisition time for a total of 16224 images was about 100 minutes, i.e. almost 10,000 images / hour. A large number of images was chosen to allow a detailed analysis of the homogeneity of the biosensor. For the main outcome of interest - the concentration of the two biomarkers, a much smaller number of images would be sufficient; for example the acquisition of 3×3 images would only take about 5:20 minutes. The magnification of the microscope objective and the size of the camera sensor yield a corresponding 208×284 μm on the biosensor. 2 Image of the field of view; after binning there are 1504×2056 pixels and the image scale is 0.138μm per pixel.
[0169] Analysis of the images was performed in parallel with data acquisition on the same computer that controlled the rest of the platform. Whenever all 169 images of a well were acquired, they were analyzed in parallel using 11 threads of the computer. Because the RGB format of the images does not provide independent brightness and color values, such normalized values were first calculated:
[0170] Brightness = sum of RGB values, normalized to the range 0...1.
[0171] Relative component = each RGB component divided by the sum of the RGB values; also normalized to the range 0...1.
[0172] The localization of particles was performed as described above. A grayscale pattern consisting of the sum of two two-dimensional Gaussian functions was used to match the donut shape of the emission from individual gold nanoparticles; the FWHM (full width at half maximum) of the pattern corresponded to an apparent particle size of 0.7 μm. A correlation of at least 0.6 between the pattern and the image was used as an acceptance criterion for potential particles; a lower threshold of 0.03 relative brightness was set to locate even potential particles with less than 10% of the brightness of the GNPs.
[0173] To obtain the average values of the intensity and emission spectrum of each particle, the previously calculated image layers (normalized brightness, normalized relative color components) are smoothed with a Gaussian filter with a σ of 2.5 pixels, which corresponds to averaging over an image area with a FWHM of approximately 0.8 μm (i.e., close to a typical apparent particle size). The filter layers are evaluated at the positions obtained in the localization step, resulting in a list of characteristics of each particle (one list per image).
[0174] Based on the characteristics obtained in the previous step, the normalized brightness and one color component are selected to calculate a two-dimensional histogram based on all the particles found. The two most prominent "peaks" correspond to monomers with particles of 80 nm (M80) and 100 nm (M100). The corresponding dimers (D80, D100) can also be easily distinguished, while larger agglomerates essentially overlap (see Figure 5 ).
[0175] At low brightness values (<0.1), particles classified as "noise" can be seen; residues from the manufacturing process can be distinguished due to their different emission spectra (<0.33); according to their brightness, they are classified into two categories (0.1 ... 0.7 low; >0.7 high). Based on this histogram, classification rules are defined. The classification rules are applied to the results from the characterization step so that the particle class is added to the already known properties of each particle found.
[0176] Based on the classification obtained in the previous step, the particles in each class are counted in each image. The results can be visualized as a table with one row per image and one column per particle class. The results per particle class are also visualized as a "heat map", i.e., the number of particles is displayed with a position similar to that on a biosensor (see Figure 6 )'s subplots (axes) in color-coded or grayscale form.
[0177] Typically, more than one image per well (13×13 images in this experiment) is taken to improve the counting statistics. Here, the outer two rows and two columns of the image of the well are within the non-sensitive area of the biosensor. The remaining 9×9=81 images are within the sensitive area and are used to calculate the main result for that well - the mean plus its coefficient of variation (CV). In this experiment, the average number of particles was calculated after applying a 40% trimming to the 81 images.
[0178] Once a calibration curve is obtained that relates, for example, the number of monomers of one type of GNP to the concentration of the biomarker of interest in the sample, particle counts can be converted to biomarker concentrations (see Table 4).
[0179] Table 4
[0180]
[0181] As used herein, the term "comprises" and its derivatives (eg, "comprising", etc.) should not be understood in an exclusive sense, that is, these terms should not be interpreted as excluding the possibility that the things described and defined may include additional elements, steps, etc.
[0182] On the other hand, the present invention is obviously not limited to the specific embodiments described herein, but also includes any variation (for example, with regard to the choice of materials, dimensions, components, configurations, etc.) that may be considered by a person skilled in the art to be within the overall scope of the invention as defined in the claims.
Claims
1. A biosensing detection platform for simultaneous, multiplexed, high-throughput and ultrasensitive optical detection of biomarkers labeled with plasmonic nanoparticles, the platform being provided with: A biosensor comprising a substrate functionalized with a recognition element that binds to the biomarker, Broadband continuous spectrum lighting system, A micrometer-scale dense array of spectrometers, Autofocus system, and an optical system adapted to collect scattered light from the surface of the biosensor onto a micrometer-sized dense array of the spectrometer to analyze different spatial locations of a sample, At the same time, each spectrometer in the array allows spectral analysis of each point of the sample, and processor, The platform is provided with translation means for moving the optical system and / or the biosensor, such that the optical system and the biosensor can be moved relative to each other in three dimensions, and wherein, The processor: i) capturing spatial and spectral scattering signals and thereby obtaining spatially and spectrally resolved images; ii) correcting the image to correct black level, brightness non-uniformity and reduce noise; iii) locating potential particles in the image by calculating a normalized two-dimensional cross-correlation between the image and a test pattern; iv) identifying particles in the image and characterizing the particles by at least color and brightness of the particles; v) classifying the particle to determine whether the particle is a labeled plasmonic nanoparticle; vi) counting the nanoparticles; vii) calculating the biomarker concentration by the quantity of said nanoparticles according to the previous calibration.
2. The biosensor detection platform according to claim 1, wherein: The light source of the illumination system is a tungsten lamp, a halogen lamp, a xenon lamp or a continuous laser or a combination of multiple LEDs or lasers in the visible and near infrared spectral range.
3. The biosensor detection platform according to claim 1 or 2, wherein: The optical system includes a microscope objective.
4. The biosensor detection platform according to claim 1 or 2, wherein: The optical system includes an infinity-corrected microscope objective combined with a tube lens.
5. The biosensor detection platform according to claim 1 or 2, wherein: The autofocus system is based on a programming device.
6. The biosensor detection platform according to claim 1 or 2, wherein: The micron-scale dense array of the spectrometer is a dense array of multiple monochrome CMOS sensors or CCD sensors coupled with a dichroic filter, or a vertically stacked array of photodetectors capable of detecting different spectral bands at the same spatial location.
7. The biosensor detection platform according to claim 1 or 2, wherein: The spectrometer's micrometer-sized dense array is a CMOS sensor or CCD sensor coupled to an array of filters arranged in a mosaic pattern.
8. The biosensor detection platform according to claim 7, wherein: The filter is an RGB filter.
9. An optical detection method, which uses the biosensor detection platform according to any one of claims 1 to 8 to perform optical detection on multiple protein biomarkers, wherein: The method comprises the following steps: a. Capturing spatial and spectral scattering signals and thereby obtaining spatially and spectrally resolved images; b. Correcting the image to correct black level, brightness non-uniformity and reduce noise; c. locating potential particles in the image by calculating a normalized two-dimensional cross-correlation between the image and a test pattern; d. identifying particles in the image and characterizing the particles at least by color and brightness of the particles; e. classifying the particles to determine whether the particles are labeled plasmonic nanoparticles; f. counting the nanoparticles; g. Calculating the biomarker concentration according to the number of nanoparticles based on the previous calibration.
Citation Information
Patent Citations
Liquid phase chip detector based on quantum dot
CN101988897A
Dual-detection biochemical sensing detector integrated with optofluidics
CN102680452A
Analyte detection enhancement by targeted immobilization, surface amplification, and pixelated reading and analysis
CN105358979A
Optical detector
CN107003122A
Ultra-wide spectrum multichannel laser flow cytometer
CN107290265A
