Method and system for three-dimensional light sheet imaging

Abstract

This paper discloses a light-sheet imaging system for imaging fluorescent samples. It also discloses a sample holder system for high-throughput light-sheet imaging of multiple three-dimensional samples without user intervention. Furthermore, this paper discloses an automated image processing method for identifying and quantifying fluorescent particles within a three-dimensional image set without user intervention or user bias.

Description

[0001] Cross-references

[0002] This application claims the benefit of U.S. Provisional Application No. 62 / 946,373, filed December 10, 2019, and U.S. Provisional Application No. 63 / 082,219, filed September 23, 2020, the entire contents of which are hereby incorporated herein by reference. Background Technology

[0003] Light sheet imaging is used to illuminate thin sections of a sample. Light sheet imaging techniques offer a faster image acquisition rate than comparable point-scanning methods such as laser confocal scanning microscopy. Improvements in light sheet imaging techniques are needed to increase sample throughput and reduce the need for user intervention. Summary of the Invention

[0004] In various aspects, this disclosure provides a light sheet imaging system comprising: a collimating illumination source configured to emit a collimated beam along a beam path; a Powell lens positioned on the beam path after the collimating illumination source and configured to extend the collimated beam along a single axis; a first cylindrical lens positioned on the beam path after the Powell lens and configured to re-collimate the beam along the single axis to produce a collimated beam having a minor axis orthogonal to the propagation direction and a major axis orthogonal to both the minor axis and the propagation direction; and a second cylindrical lens positioned on the beam path after the second cylindrical lens and configured to focus the beam along the minor axis onto a focal plane.

[0005] In some aspects, the three-dimensional sample is positioned at the focal plane. In some aspects, the three-dimensional sample comprises fluorescent droplets and / or fluorescently labeled cells. In some aspects, the light sheet imaging system further includes a camera comprising an imaging path orthogonal to the beam path, wherein the camera is positioned such that the imaging path intersects the beam path at the focal plane.

[0006] In various aspects, this disclosure provides a sample holder comprising: a fluid reservoir configured to contain fluid, the fluid reservoir including a base, a first transparent surface, a second transparent surface connected at right angles to the first transparent surface, and a third transparent surface opposite to and connected to the first and second transparent surfaces; a tube rack configured to hold a row of tubes partially immersed in the fluid; and a translation stage supporting the tube rack and configured to translate the tube rack along a translation axis oriented at a 45° angle relative to the first and second transparent surfaces and parallel to the row of tubes.

[0007] In some aspects, the sample holder further includes a lid configured to fit on top of the first, second, and third transparent surfaces and configured to retain fluid. In some aspects, the lid further includes an O-ring positioned at the contact point between the lid and the first, second, and third transparent surfaces, wherein the O-ring is configured to prevent fluid leakage. In some aspects, the lid includes an opening configured to receive a tube holder and a translation stage. In some aspects, the sample holder further includes a first opaque surface and a second opaque surface, the first opaque surface connecting the first transparent surface to the third transparent surface, and the second opaque surface connecting the second and third transparent surfaces.

[0008] In various aspects, this disclosure provides a sample holder comprising: a fluid reservoir configured to contain fluid, the fluid reservoir including a base, a first transparent surface, a second transparent surface connected at right angles to the first transparent surface, and a third transparent surface opposite to and connected to the first and second transparent surfaces; a tube rack configured to hold a row of tubes partially immersed in the fluid; and a translation stage supporting the fluid reservoir and the tube rack and configured to translate the fluid reservoir and the tube rack along a first translation axis and a second translation axis, wherein the first translation axis is parallel to the first transparent surface and oriented perpendicular to the second transparent surface, and wherein the second translation axis is perpendicular to the first transparent surface and parallel to the second transparent surface, and wherein the first and second translation axes are oriented at a 45° angle relative to the row of tubes.

[0009] In some aspects, the sample holder further includes a fourth transparent surface and a fifth transparent surface, the fourth transparent surface being connected at a right angle to the second transparent surface and parallel to the first transparent surface, and the fifth transparent surface being connected at a right angle to the fourth transparent surface and parallel to the second transparent surface. In some aspects, the sample holder further includes a plurality of transparent surfaces connected in series at right angles to the second transparent surface, wherein alternating transparent surfaces of the plurality of transparent surfaces are parallel to the first transparent surface or parallel to the second transparent surface. In some aspects, the sample holder further includes a lid configured to fit on top of the first, second, and third transparent surfaces and configured to retain fluid. In some aspects, the lid further includes an O-ring positioned at the contact point between the lid and the first, second, and third transparent surfaces, wherein the O-ring is configured to prevent fluid leakage.

[0010] In some aspects, the fluid includes a refractive index-matching fluid. In some aspects, the sample rack comprises a polymer. In some aspects, the multiple tubes include PCR tubes. In some aspects, the multiple tubes include microcentrifuge tubes. In some aspects, the sample rack is configured to hold up to 12 tubes. In some aspects, the sample rack is configured to hold up to 8 tubes. In some aspects, the sample rack is configured to hold up to 4 tubes.

[0011] In various aspects, this disclosure provides an image processing method comprising: collecting a three-dimensional image dataset including multiple cross-sectional images; identifying multiple substantially spherical fluorescent particles in the three-dimensional image dataset; dividing one or more of the multiple cross-sectional images into multiple image grids; independently applying an intensity threshold to one or more of the multiple image grids; and identifying one or more signal-positive particles from the multiple substantially spherical fluorescent particles, wherein the one or more signal-positive particles have an intensity greater than the intensity threshold.

[0012] In some aspects, the method further includes applying a smoothing filter to one or more cross-sectional images of a plurality of cross-sectional images before identifying a plurality of substantially spherical fluorescent particles. In some aspects, the method further includes applying a median filter to one or more cross-sectional images of a plurality of cross-sectional images before identifying a plurality of substantially spherical fluorescent particles. In some aspects, identifying a plurality of substantially spherical fluorescent particles includes performing a three-dimensional convolution using a three-dimensional template.

[0013] In some aspects, the method further includes determining the intensity of one or more substantially spherical fluorescent particles among a plurality of substantially spherical fluorescent particles through three-dimensional local maximum analysis. In some aspects, the method further includes performing a second three-dimensional maximum analysis to remove one or more asymmetric particles from the plurality of substantially spherical fluorescent particles. In some aspects, the intensity threshold is determined by logarithm of the intensity of the image grid. 10 The median of the histogram is determined by adding or subtracting the standard deviation. In some respects, the intensity threshold maximizes the interval between one or more intensities of the positive particles and the intensity threshold.

[0014] In some aspects, the method further includes repeatedly applying an intensity threshold to one or more image grids corresponding to ratio outliers, and determining a refinement ratio of the one or more image grids corresponding to the ratio outliers after applying the intensity threshold. In some aspects, the method further includes selecting a region of interest in one or more cross-sectional images among a plurality of cross-sectional images, wherein the region of interest includes a sample region. In some aspects, the method further includes identifying a background intensity and subtracting the background intensity from the intensity of one or more cross-sectional images among a plurality of cross-sectional images. In some aspects, the method further includes selecting a subset of a plurality of substantially spherical fluorescent particles, wherein the substantially spherical fluorescent particles comprise sizes greater than a minimum size threshold and less than a maximum size threshold. In some aspects, the method further includes identifying a plurality of signal-positive particles comprising intensities above a minimum threshold from the plurality of substantially spherical fluorescent particles.

[0015] In some aspects, the method further includes determining whether a low-signal region exists in a cross-sectional image, and if a low-signal region exists, identifying low-signal and high-signal regions within a region of interest (ROI) of one or more cross-sectional images and specifying a low-signal image grid that includes the low-signal region, or if no low-signal region exists, specifying the ROI as a high-signal region. In some aspects, the method further includes determining a ratio of the number of signal-positive particles to the number of multiple spherical fluorescent particles for one or more image grids. In some aspects, the method further includes identifying ratio outliers, which include ratios that are higher than or lower than the median ratio of the multiple image grids. In some aspects, the method further includes removing ratio outliers. In some aspects, the method is automated. In some aspects, the method is performed without user intervention. In some embodiments, the method further includes performing cluster removal to eliminate overlapping fluorescent droplets. In some embodiments, cluster removal includes eliminating droplets with a peak density below a set threshold. In some embodiments, the peak density is determined based on droplet size. In some embodiments, the droplet size is calculated based on the full width at half maximum (WHM). In some embodiments, cluster removal includes nearest neighbor / volume exclusion / non-maximum suppression. In some implementations, cluster removal includes determining the number of close droplet neighbors for each droplet. In some implementations, cluster removal also includes generating a histogram of close droplet neighbors. In some implementations, cluster removal further includes eliminating anomalous droplets with a determined number of neighbors exceeding a predetermined threshold. In some implementations, the predetermined threshold is based on the probability that a merged droplet cluster is detected as a false positive.

[0016] This document also provides a method for analyzing multiple samples, the method comprising automatically subjecting each of the multiple samples, one sample at a time, to a laser beam, thereby generating emitted light in each of the multiple samples, wherein the multiple samples include at least four samples. In some aspects, the multiple samples are located in a sample holder. In some aspects, the sample holder is moved in a direction perpendicular to the illumination path of the laser beam to remove a scanned first sample from the laser beam and move a second sample into the beam for scanning. In some aspects, the method further includes moving the laser beam relative to the samples while subjecting each of the multiple samples to the laser beam. In some aspects, moving the laser beam relative to the samples includes pivoting movement of the laser beam. In some aspects, moving the laser beam relative to the samples includes translational movement of the laser beam. In some aspects, pivoting or translational movement reduces artifacts in the resulting image data compared to methods without pivoting or translational movement. In some aspects, the laser beam is moved relative to the samples at a frequency of at least about 0.1 kHz. In some aspects, the laser beam is moved relative to the samples at a frequency of about 0.1 kHz to about 20 kHz. In some aspects, the laser beam is moved relative to the samples at a frequency of at least about 1 kHz. In some aspects, the laser beam is a laser sheet. In some aspects, the plurality of samples includes at least eight samples. In some aspects, the plurality of samples are subjected to a second laser beam. In some aspects, the second laser beam is configured to irradiate a portion of one of the plurality of samples opposite to the portion irradiated by the first laser beam. In some aspects, each of the plurality of samples is subjected to both the first laser beam and the second laser beam simultaneously.

[0017] Incorporation

[0018] All publications, patents and patent applications mentioned in this specification are incorporated herein by reference to the extent that each individual publication, patent or patent application is specifically and individually cited and incorporated herein by reference. Attached Figure Description

[0019] The patent or application documents contain at least one color drawing. The patent office will, upon request and upon payment of the necessary fees, provide a copy of the patent or patent application publication accompanied by the color drawing. The novel features of this disclosure are specifically set forth in the appended claims. A better understanding of the features and advantages of this disclosure will be obtained by referring to the following detailed description and drawings, in which illustrative embodiments utilizing the principles of this disclosure are set forth, in which:

[0020] Figure 1 The diagram shows the localization of the sample within the imaging field of view under a light sheet imaging system (left) and different cross-sectional depths (right, (i), (ii) and (iii));

[0021] Figure 2An exemplary sample holder design is shown for sequential imaging of multiple tubes without user intervention using light sheet imaging. Figure 2 A shows an isometric view of the sample holder. Figure 2 B shows an image of the sample holder and sample tube positioned within the light sheet imaging system;

[0022] Figure 3 An exemplary sample holder design is shown for sequential imaging of multiple tubes without user intervention using light sheet imaging. Figure 3 A shows an isometric view of the sample holder. Figure 3 B shows a top view of the sample holder within the light sheet imaging system;

[0023] Figure 4 An exemplary optical configuration for generating a focusing plate from a collimated laser source is shown. Figure 4 A illustrates an optical configuration comprising three cylindrical lenses and one aspherical lens. Figure 4 B illustrates an optical configuration including one Powell lens and two cylindrical lenses;

[0024] Figure 5A , Figure 5B , Figure 5C , Figure 5D , Figure 5E , Figure 5F and Figure 5G An image processing method for identifying and counting positive and negative fluorescent particles is shown. Figure 5A The diagram illustrates region of interest (ROI) extraction used to identify regions of interest that include the sample area. Figure 5B This shows the local background subtraction from the region of interest. Figure 5C A three-dimensional (3D) convolution with a spherical template is shown for identifying essentially spherical fluorescent particles. Figure 5D The local maxima and symmetry checks for eliminating asymmetric particles using 3D convolution are shown. Figure 5E The detection of low-intensity regions (“blurred layers”) within the region of interest is shown. Figure 5F The local brightness histogram and automatic thresholding of a single image grid within the region of interest are shown. Figure 5G A global correction is shown, which is used to remove grid outliers from two distinct positive particle clusters that have a significantly higher or lower ratio of signal positive particles to total particles than the median ratio of all grids.

[0025] Figure 6 The setup of the imaging system and the method for analyzing eight or more samples using light-sheet imaging are shown.

[0026] Figure 6A shows an analyte sample (here, a 96-well plate) loaded onto the sample loading block of the imaging system.

[0027] Figure 6 B shows an 8-sample strip of a 96-well plate picked up in order to image 8 samples at once;

[0028] Figure 6 C shows eight sample strips located in the imaging chamber and scanned using an optical configuration to generate a focused light sheet from a collimated laser source;

[0029] Figure 7 Different methods for improving image quality and reducing artifacts in samples analyzed using light sheet imaging are shown.

[0030] Figure 7 A shows that pivoting relative to the laser beam along the focal plane of the sample can reduce artifacts in the resulting image;

[0031] Figure 7 B shows that translating the laser beam along the focal plane of the sample can reduce artifacts in the resulting image;

[0032] Figure 7 C shows that double-sided illumination of the sample can also reduce banding and / or shadow artifacts in the resulting image.

[0033] Figure 8A An exemplary image of droplet candidate extraction is shown;

[0034] Figure 8B An example diagram of a convolutional neural network (CNN) is shown;

[0035] Figure 8C Example plots of nonmaximum suppression and cluster removal are shown; and

[0036] Figure 9 A non-limiting example of a computing device is shown; in this case, the device has one or more processors, memory, storage, and network interfaces. Detailed Implementation

[0037] Three-dimensional imaging can be used to image three-dimensional samples such as organisms, tissues, cells, or liquid samples. Three-dimensional imaging can achieve higher throughput imaging compared to comparable one-dimensional or two-dimensional imaging techniques by imaging larger sample volumes than one-dimensional or two-dimensional imaging techniques. Three-dimensional imaging techniques, such as light-sheet imaging, can be used to image fluorescent samples containing multiple fluorescent particles present in different cross-sectional planes of the sample. In some embodiments, fluorescent samples may include sample tubes containing fluorescent particles or droplets, cells with labeled proteins or nucleic acids, immunofluorescent tissue samples, or fluorescently labeled organisms. For example, a three-dimensional fluorescent sample may be a digital polymerase chain reaction (dPCR) sample containing both positive and negative fluorescent droplets.

[0038] Compared to one-dimensional and two-dimensional imaging techniques, three-dimensional imaging techniques offer many advantages, such as higher throughput imaging methods and sequential imaging of multiple three-dimensional samples without user intervention. However, compared to one-dimensional and two-dimensional imaging techniques, three-dimensional imaging techniques may have additional complexities due to the nature of manipulating three-dimensional samples and processing three-dimensional image datasets, such as non-uniform background intensity, non-uniform illumination intensity, or signal variability at different locations on or within different cross-sectional planes.

[0039] This document discloses a light-sheet imaging system for imaging fluorescent samples. The light-sheet imaging system may include an illumination system configured to generate a light sheet focused onto the sample. The sample may be a three-dimensional (3D) sample, such as a sample in a test tube or vial. The sample may be a two-dimensional (2D) sample, such as a sample in a planar array or plate. The sample may be a one-dimensional (1D) sample, such as a sample in a flow channel. This document also discloses a sample holder system for high-throughput light-sheet imaging of multiple three-dimensional samples. These sample holder systems can be used without user intervention. This document further discloses an automated image processing method for identifying and quantifying fluorescent particles within a three-dimensional image set without user intervention or user bias. In some embodiments, the three-dimensional image set can be collected using light-sheet imaging of a three-dimensional sample. As used herein, fluorescent particles may be particles such as nanoparticles or beads having a fluorescent portion, droplets such as lipid droplets containing a fluorescent portion, cells containing a fluorescent portion, or molecules or portions such as organic fluorophores.

[0040] Light sheet imaging system

[0041] This paper discloses a light sheet illumination system for uniform light sheet illumination of a sample cross-sectional plane. It also discloses a sample holder for positioning and repositioning multiple three-dimensional samples within an imaging system for continuous imaging without user intervention.

[0042] Light sheet illumination system

[0043] The illumination system disclosed herein can convert a collimated illumination source into a light sheet to image a cross-sectional plane within a sample. The collimated illumination source can be a laser. In some embodiments, the light sheet can be focused along a single axis. The illumination system can be configured such that the focal plane lies within the sample. The sample can be a one-dimensional, two-dimensional, or three-dimensional sample.

[0044] exist Figure 4 An exemplary light sheet illumination configuration is shown in the figure. Figure 4 In the first configuration shown in A, an illumination source (e.g., a laser beam) passes through a collimator to produce a collimated beam. The collimated beam can pass through a first cylindrical lens, which causes the beam to converge along a first axis (e.g., a horizontal axis). The diverging beam can pass through an aspherical lens, which causes the beam to be collimated along the first axis and converge along a second axis perpendicular to the first axis (e.g., a vertical axis). The beam can then pass through a second cylindrical lens having a surface oriented perpendicular to the surface of the first cylindrical lens. The second cylindrical lens collimates the beam along the second axis, thereby producing a collimated, elliptical beam elongated along the second axis. The resulting beam may have a Gaussian profile. The beam can then pass through a third cylindrical lens having a surface oriented parallel to the surface of the second cylindrical lens. The third cylindrical lens focuses the elliptical beam along the first axis toward the focal plane. In some embodiments, the order of the optical elements can be rearranged. For example, an aspherical lens can be positioned in the beam path before the first cylindrical lens, or an aspherical lens can be positioned after the second cylindrical lens.

[0045] exist Figure 4 In the second configuration shown in B, a collimated illumination source emits a collimated beam (e.g., a laser beam). The collimated beam can pass through a line-generating lens. In some embodiments, the line-generating lens can be a Powell lens or a laser line generator lens. After passing through the line-generating lens, the beam can have a linear profile diverging along a first axis (e.g., the vertical axis). The beam can then pass through a cylindrical lens that can focus the beam along a second axis (e.g., the horizontal axis) toward the focal plane.

[0046] and Figure 4 Compared to the elliptical beam generated after the first cylindrical lens, aspherical lens, and third cylindrical lens shown in Figure A. Figure 4 The linear beam profile produced after the online generation lens shown in B may be longer and have a more uniform intensity across its height. Furthermore, Figure 4 The optical configuration shown in B is... Figure 4The optical configuration shown in A produces less spherical aberration because it uses fewer optical elements. In both configurations, long focal length lenses can be used to reduce spherical aberration.

[0047] A light-sheet imaging system may include an illumination system and a detector as described herein. The detector may be a camera. For example, the camera may be a wide-field-of-view camera. The camera may have a maximum scan area of ​​20 μL. The camera may have a maximum scan area of ​​50 μL or greater. Exemplary cameras include, but are not limited to, the QHY174 camera with a scan area of ​​1920 x 1200 pixels, 5.6 x 3.5 mm, and a resolution of 2.93 μm / pixel, and the Basler aca-2440-35μm with a scan area of ​​2448 x 2048 pixels, 8.4 x 7.1 mm, and a resolution of 3.45 μm / pixel. Cameras with larger scan areas can scan the entire cross-section of a PCR tube without interleaving. The camera may have an imaging path. The camera may be located within the light-sheet imaging system such that the imaging path is orthogonal to the illumination beam path and intersects the illumination beam path at the focal plane of the light sheet. The sample may be located at the intersection of the imaging path and the illumination beam path.

[0048] Light-sheet imaging systems can be configured to allow scanning of 1, 2, 4, 6, 8, 10, 12, 20, or more samples simultaneously without user intervention. Such imaging systems allow for high-throughput scanning of multiple samples. These samples may include digital PCR samples. A high-throughput imaging system may include multiple components, including a sample loading block, a sample holder system, an operating arm configured to move a number of samples from the sample loading block to the sample holder system, a laser source configured to provide a light sheet, and a detector configured to detect radiation (e.g., fluorescence) emitted from the samples. The sample holder block may be loaded with multiple samples (e.g., digital PCR samples), for example, such as... Figure 6 As shown in A. In some cases, multiple samples are located in the well plate, for example, a 96-well plate. The operating arm of the light sheet imaging system can be configured to hold a certain number of sample tubes, for example, about 1, 2, 4, 6, 8, 10, 12, 20 or more sample tubes, and as shown in Figure A. Figure 6 The eight sample strips shown in B are placed within the sample holder system described herein. The sample holder system, including the samples, can be configured to move along an axis such that each sample located in the holder can be scanned by a light source and the radiation emitted by it can be detected in a detector unit.

[0049] In some cases, the sample holder system can be configured to provide additional movement for the sample, such as pivoting and / or translating relative to the light sheet. In various embodiments, the light sheet can be moved while the sample is being scanned. In such cases, the sample can remain stationary during scanning, while the light sheet moves relative to the sample. Such movement of the light sheet can include pivoting and / or translating movement. The movement frequency of the light sheet can be faster than the exposure time of the sample. In such cases, the light sheet may move at at least about 0.1, 0.5, 0.7, 1.0, 1.5, 2, 2.5, 5, 10, 15, 20 kHz (10 3 s -1 The sheet may move at a frequency of at least about 1 kHz or higher, or any frequency in between. In such cases, the sheet may move at frequencies ranging from 0.1 to 20 kHz, 0.1 to 30 kHz, 0.5 to 20 kHz, 0.7 to 20 kHz, 1 to 20 kHz, 2 to 20 kHz, 5 to 20 kHz, 10 to 20 kHz, 15 to 20 kHz, 0.1 to 1 kHz, 0.1 to 2 kHz, 0.1 to 5 kHz, 0.1 to 10 kHz, 0.1 to 15 kHz, 0.5 to 5 kHz, 1 to 10 kHz, or 0.5 to 2 kHz. In some cases, such rapid pivoting of the sheet can be achieved, for example, by using a resonant mirror or a galvanometer resonant scanner. Figure 7 A- Figure 7 As shown in B, this additional movement of the light sheet can reduce artifacts in the resulting image and thus improve image quality. Such artifacts may be caused by light scattering and absorption events during the scanning of the sample using the light sheet.

[0050] In some implementations, the imaging system is configured to provide double-sided illumination of the sample, such as Figure 7 As shown in C. This setup can also improve image quality by reducing artifacts, where light from each light source only needs to pass through half of the sample to illuminate the entire sample cross-section.

[0051] Sample rack system

[0052] The sample holder system described herein can be used to hold multiple three-dimensional samples for continuous imaging without user intervention. The sample holder may include open or closed chambers. Three-dimensional samples may be tubes containing liquid, aqueous, or gel samples. Three-dimensional samples may be tissue samples in a tissue sample holder. Three-dimensional samples may be biological organisms.

[0053] The sample holder design disclosed herein allows for automated imaging of multiple samples (e.g., tubes arranged in a strip tube format) without user intervention. In some embodiments, samples can be imaged using three-dimensional scanning imaging methods. For example, samples can be imaged using the light sheet imaging disclosed herein, or using confocal imaging. The sample holder design also allows for a single scan of samples with large cross-sections (e.g., 50 μL PCR tubes) without the need for multiple images to be staggered across each cross-section. The sample holder can be configured to accommodate up to 2, 3, 4, 5, 6, 7, 8, 9, 10, 12, 14, 15, 16, 18, 20, 24, or more three-dimensional samples to be imaged without user intervention.

[0054] Figure 1 and Figure 2 The first sample holder design shown may include a fluid reservoir. For example... Figure 2 As shown in A, the fluid reservoir may be elongated to accommodate translation of a sample partially immersed in the fluid within the reservoir. The fluid reservoir may include three transparent surfaces (“transparent surfaces” indicated by arrows). Two transparent surfaces—a first transparent surface and a second transparent surface—may be connected at right angles. The first transparent surface may be configured to transmit illumination light (e.g., a light sheet). The second transparent surface may be configured to transmit emitted light (e.g., fluorescence emission) toward a detector (e.g., a camera). Figure 2 The third transparent surface, located behind the first and second transparent surfaces as shown in Figure A, can be configured to transmit illumination light out of the fluid reservoir after the illumination light has passed through the sample, thereby reducing reflection within the sample chamber. The remaining surfaces of the fluid reservoir can be opaque or partially opaque to reduce reflection within the fluid reservoir.

[0055] like Figure 2 As shown in B, a sample, such as multiple tubes, can be positioned in a sample holder. The sample holder can be supported by a translation stage. The translation stage can be configured to translate the sample along a translation axis when the sample is partially immersed in a fluid. The translation axis can be parallel to the long axis of the fluid reservoir. The translation axis can be at a 45° angle relative to the surfaces of the first and second transparent surfaces. Optionally, the translation axis can be parallel to the surface of the third transparent surface. In some embodiments, the translation stage can be configured to translate the sample holder along a first translation axis and a second translation axis. The first translation axis can be parallel to the surface of the first transparent surface. The second translation axis can be parallel to the surface of the second transparent surface. Optionally, the first and second translation axes can be parallel to the surface of the third transparent surface. The fluid reservoir can contain a refractive index matching liquid. The refractive index matching liquid can match the refractive index of the sample.

[0056] The fluid reservoir and sample holder can be positioned within a three-dimensional imaging system. For example, the fluid reservoir and sample holder can be positioned within a light sheet imaging system as described herein. The sample holder can be positioned such that a first sample is positioned at the intersection of the illumination beam path and the imaging path, and translation of the sample along a translation axis enables imaging of continuous cross-sectional planes within the first sample, such as... Figure 1 As shown in the diagram. The cross-sectional image plane within the sample can be offset relative to the camera's field of view depending on the position of the cross-sectional plane within the sample, such as... Figure 1 As shown in diagrams (i), (ii), and (iii). In some embodiments, the cross-sectional image plane within the sample is not offset relative to the camera's field of view based on the position of the cross-sectional plane within the sample. Further translation along the translation axis can position the second sample at the intersection of the illumination beam path and the imaging path.

[0057] Figure 3 The second sample holder design shown may include a fluid reservoir. The fluid reservoir can be configured to contain and hold, for example... Figure 3 A sample holder, as shown in Figure A, contains multiple samples partially immersed in a fluid within a fluid reservoir. The fluid reservoir may include multiple transparent surfaces (represented by arrows as "transparent surfaces"). Pairs of transparent surfaces may be connected at right angles, such that the multiple transparent surfaces are arranged in a zigzag pattern. The number of pairs of transparent surfaces may correspond to the number of samples that the sample holder can hold. The first transparent surface of each pair may be configured to transmit illumination light (e.g., a light sheet). The second transparent surface of each pair may be configured to transmit emitted light (e.g., fluorescence emission) toward a detector (e.g., a camera). Figure 3 The additional transparent surface shown in A, located behind the multiple transparent surfaces, can be configured to transmit illumination light out of the fluid reservoir after it has passed through the sample, thereby reducing reflections within the sample chamber. The remaining surfaces of the fluid reservoir can be opaque or partially opaque to reduce reflections within the fluid reservoir.

[0058] like Figure 2As shown in Figure B, a sample, such as multiple tubes, can be positioned in a sample holder. The sample can be positioned such that the illumination beam is not obstructed when entering and exiting the sample. The sample holder and fluid reservoir can be supported by a translation stage. The translation stage can be configured to translate the sample and fluid reservoir along a first translation axis and a second translation axis. The first translation axis can be parallel to the surface of a first transparent surface of each pair. The second translation axis can be parallel to the surface of a second transparent surface of each pair. Optionally, the first and second translation axes can be parallel to the surface of an additional transparent surface. The translation stage can be configured to translate the sample and fluid reservoir along the translation axes while the sample is partially immersed in the fluid. The translation axes can be at a 45° angle relative to the surfaces of the first and second transparent surfaces of each pair. Optionally, the translation axes can be parallel to the surface of an additional transparent surface. The fluid reservoir can contain a refractive index matching liquid. The refractive index matching liquid can match the refractive index of the sample. The fluid reservoir can also include a cap. The cap can include an O-ring to prevent fluid leakage.

[0059] The fluid reservoir and sample holder can be positioned within a three-dimensional imaging system. For example, the fluid reservoir and sample holder can be positioned within a light sheet imaging system as described herein. The sample holder can be positioned such that a first sample is positioned at the intersection of the illumination beam path and the imaging path, with the illumination beam path passing through a first pair of first transparent surfaces and the imaging path passing through a second transparent surface of the first pair. Translation of the sample along a first translation axis enables imaging of continuous cross-sectional planes within the first sample, such as... Figure 3 As shown in B. Translation along the second translation axis positions the second sample at the intersection of the illumination beam path and the imaging path, and positions the second pair of transparent surfaces within the illumination beam path and the imaging path. Compared to the first sample holder design, the second sample holder design reduces fluid evaporation and wicking, provides a smaller image and instrument footprint, and reduces the complexity of image analysis.

[0060] Image processing methods for fluorescence particle quantization

[0061] The image processing method described herein can be used to identify and quantify fluorescent particles within a three-dimensional image dataset. Fluorescent particles can be droplets (e.g., droplets containing nucleic acid molecules, protein molecules, or cells). Fluorescent particles can be particles or cells suspended in a solution. Fluorescent particles can be particles, molecules, or cells suspended in a gel matrix.

[0062] The resulting cross-sectional planar images can be combined to generate a 3D image dataset for each sample. Regions of interest within each cross-section can be selected to delineate the sample boundaries. Figure 5A This ensures that only fluorescent particles within the tube cross-section are considered. Local background subtraction can be performed to suppress background fluorescence. Figure 5BA smoothing filter can be applied to smooth the image, and a medium filter can be applied to remove hot pixels. In some implementations, the smoothing filter can be a Gaussian filter.

[0063] Three-dimensional (3D) convolution can be performed using a 3D template to identify fluorescent particles (e.g., corresponding to fluorescent lipid droplets) that substantially conform to the desired shape and are within the expected size range, and to eliminate fluorescent particles that do not conform to the expected size and shape parameters. Figure 5C In some implementations, the three-dimensional template can be a spherical template, an elliptical template, or any other three-dimensional shape. 3D local maxima analysis can be performed to identify potential fluorescent particle candidates with positive fluorescence signals. Figure 5D Optionally, additional 3D local maximum analysis can be performed to filter out asymmetric particle candidates (e.g., non-spherical particle candidates).

[0064] The boundary between the bulk sample and the top layer of the sample (called the "blurred layer"), which has a lower signal than the bulk sample due to refractive index mismatch at the air-sample interface, can be defined. Figure 5E The region above the blur layer boundary can be grouped into a grid for separate local histogram analysis from the grid of the image containing large samples. Each tube cross-section can be divided into multiple grids for individual thresholding. Figure 5F For each grid cell, the logarithmic intensity of each fluorescent particle candidate can be plotted. 10 (log 10 A histogram of intensity. The threshold log for each grid cell can be automatically set. 10 (Intensity) values ​​are used to distinguish noise from signal. A threshold can be set based on the following assumptions: (A) most particles belong to a population with some outliers (e.g., a population of lipid droplets that are mostly negative for the signal, or a population of lipid droplets that are mostly positive for the signal), or (B) two populations of particles exist. For case (A), a median plus or minus a standard deviation threshold can be used to automatically separate positive and negative signal particles. For case (B), an optimization algorithm can be used to select a threshold that maximizes the separation between the positive signal population and the threshold, such as by measuring the number of standard deviations of the positive population above the threshold.

[0065] The ratio of positive to negative particles can be determined for each grid cell. It can be expected that, despite differences in brightness, the ratio of positive to negative particles should not vary significantly between grid cells. A large deviation of the ratio in a given grid cell from the median ratio may indicate an inappropriate threshold selection. Grid cells with ratios significantly higher or lower than the median ratio can be identified as outliers and can be automatically selected for an additional round of refinement and thresholding detection of one or more (e.g., two) distinct clusters of positive particles being counted. Figure 5G After further refinement, the remaining outliers can be removed, and the global positive particle count can be determined.

[0066] In some implementations, such as Figures 8A to 8C As shown, a convolutional neural network (CNN) is used to distinguish noise and signal, either as an adjunct to or alternative to a distinction based on an intensity-based cutoff value. In some implementations, the CNN learns to distinguish positive and negative droplets based on the 3D intensity distribution of the fluorescent droplets as a whole (which includes information about the droplet's shape and size), rather than using the maximum pixel intensity of each droplet and an intensity-based histogram cutoff value.

[0067] In some implementations, according to Figure 8A The CNN identifies fluorescent droplet candidates based on 3D local maxima analysis. In some implementations, the CNN is trained using synthetic data, real-world datasets, or both at various positive droplet occupancy rates and signal-to-noise ratios (SNR). In some implementations, hard negative data is used to additionally or alternatively train the CNN. In some implementations, training with such hard negative data improves the CNN's ability to identify ambiguous fluorescent droplet candidates.

[0068] In some embodiments, the imaging processing method of this invention further includes cluster removal to eliminate overlapping fluorescent droplets. In some embodiments, cluster removal includes eliminating droplets with a peak density below a set threshold. In some embodiments, the peak density is determined based on droplet size. In some embodiments, the droplet size is calculated based on the full width at half maximum (FWHM). In some embodiments, cluster removal includes nearest neighbor / volume exclusion / non-maximum suppression. In some embodiments, cluster removal includes determining the number of close droplet neighbors for each droplet. In some embodiments, cluster removal further includes generating a histogram of close droplet neighbors. In some embodiments, cluster removal further includes eliminating anomalous droplets with a determined number of neighbors exceeding a set threshold. In some embodiments, the set threshold is based on the probability that a merged droplet cluster is detected as a false positive. In some embodiments, such cluster removal assumes that all droplets are randomly distributed and that the density is relatively uniform across the entire tube volume. In some embodiments, the imaging processing method of this invention further includes calculating the SNR of each droplet individually and eliminating droplets with an SNR below a set threshold.

[0069] In some implementations, the image processing method described herein achieves improved detection of fluorescent particle candidates by mitigating the inherent low signal-to-noise ratio, high signal and background variations, and blurred interfaces in many 3D images of PCR droplets in tubes.

[0070] Machine Learning

[0071] In some implementations, machine learning algorithms are used to assist in the detection of fluorescent droplet candidates. In some implementations, the machine learning algorithms used for detecting fluorescent droplet candidates employ one or more forms of labeling, including but not limited to manually labeled labels and semi-supervised labels. Manually labeled labels can be provided using hand-made heuristics. Semi-supervised labels can be determined using clustering techniques to find fluorescent droplet candidates similar to those previously labeled with manually labeled and semi-supervised labels. Semi-supervised labels can employ XGBoost, neural networks, or both.

[0072] In some implementations, the training set is expanded by generating synthetic 3D images. In other implementations, synthetic 3D images are generated by approximating the droplets as spheres or spots with different 3D aspect ratios.

[0073] In some implementations, semi-supervised labeling includes augmentations applied to synthetic images, real data, or both. In some implementations, augmentations include signal strength, signal variation, SNR, polydispersion, transformation, or any combination thereof. In some implementations, transformations include dilation, expansion, reflection, rotation, shearing, stretching, translation, or any combination thereof. In some implementations, SNR augmentation adds noise (e.g., salt and pepper, stripes) to the image to make the model most robust to noise. In some implementations, such noise augmentations are annotated with ground truth based on the exact location of the generated droplets, augmentations, or both. In some implementations, increased training model size and diversity improve the robustness of the machine learning algorithms presented in this paper.

[0074] In some implementations, the machine learning algorithm used to detect fluorescent droplet candidates employs a remote supervision method. Remote supervision methods can create a large training set seeded from a small, hand-annotated training set. Remote supervision methods can include positive label-free learning using the training set as the "positive" class. Remote supervision methods can employ logistic regression models, recurrent neural networks, or both.

[0075] Examples of machine learning algorithms can include support vector machines (SVM), Naive Bayes classification, random forests, neural networks, deep learning, or other supervised or unsupervised learning algorithms for classification and regression. Machine learning algorithms can be trained using one or more training datasets. In some implementations, machine learning algorithms utilize regression modeling, where the relationship between the predictor variable and the dependent variable is determined and weighted. For example, in one implementation, fluorescent droplet candidates can be the dependent variable and are derived from an intensity-based histogram.

[0076] A non-restricted example of a multiple linear regression model algorithm is shown below: Probability = A0 + A1(X1) + A2(X2) + A3(X3) + A4(X4) + A5(X5) + A6(X6) + A7(X7)... where Ai (A1, A2, A3, A4, A5, A6, A7, ...) are the "weights" or coefficients discovered during regression modeling; and Xi (X1, X2, X3, X4, X5, X6, X7, ...) are data collected from the user. The model can contain any number of Ai and Xi variables.

[0077] Computing System

[0078] refer to Figure 9The diagram illustrates a block diagram depicting an exemplary machine, including a computer system 1300 (e.g., a processing or computing system) in which a set of instructions can be executed to cause the device to perform or implement any one or more aspects and / or methods of static code scheduling of the present disclosure. Figure 9 The components described are merely examples and do not limit the scope or functionality of any hardware, software, embedded logic components, or combinations of two or more such components used to implement a particular implementation.

[0079] Computer system 1300 may include one or more processors 1301, memory 1303, and storage 1308, which communicate with each other and with other components via bus 1340. Bus 1340 may also link a display 1332, one or more input devices 1333 (which may include, for example, a keypad, keyboard, mouse, stylus, etc.), one or more output devices 1334, one or more storage devices 1335, and various tangible storage media 1336. All these components may engage with bus 1340 directly or via one or more interfaces or adapters. For example, various tangible storage media 1336 may engage with bus 1340 via storage media interface 1326. Computer system 1300 may have any suitable physical form, including but not limited to one or more integrated circuits (ICs), printed circuit boards (PCBs), mobile handheld devices (such as mobile phones or PDAs), laptops or notebook computers, distributed computer systems, computing grids, or servers.

[0080] Computer system 1300 includes one or more processors 1301 (e.g., a central processing unit (CPU) or a general-purpose graphics processing unit (GPGPU)) that perform functions. Processor 1301 optionally includes a cache memory unit 1302 for temporary local storage of instructions, data, or computer addresses. Processor 1301 is configured to assist in the execution of computer-readable instructions. As a result of processor 1301 executing non-transitory processor-executable instructions embodied in one or more tangible computer-readable storage media such as memory 1303, storage 1308, storage device 1335, and / or storage medium 1336, computer system 1300 can be Figure 9The components depicted herein provide functionality. A computer-readable medium may store software implementing a particular implementation, and processor 1301 may execute that software. Memory 1303 may read the software from one or more other computer-readable media (such as mass storage devices 1335, 1336) or from one or more other sources via a suitable interface such as network interface 1320. The software may cause processor 1301 to perform one or more processes or steps of one or more processes described or illustrated herein. Performing such processes or steps may include defining data structures stored in memory 1303 and modifying the data structures in accordance with the guidance of the software.

[0081] Memory 1303 may include various components (e.g., machine-readable media), including but not limited to random access memory components (e.g., RAM 1304) (e.g., static RAM (SRAM), dynamic RAM (DRAM), ferroelectric random access memory (FRAM), phase-change random access memory (PRAM), etc.), read-only memory components (e.g., ROM 1305), and any combination thereof. ROM 1305 may function to communicate data and instructions unidirectionally to processor 1301, while RAM 1304 may function to communicate data and instructions bidirectionally with processor 1301. ROM 1305 and RAM 1304 may include any suitable tangible computer-readable media described below. In one example, a basic input / output system 1306 (BIOS) may be stored in memory 1303, including basic routines such as those that help pass information between elements within computer system 1300 during startup.

[0082] Fixed storage 1308 is bidirectionally connected to processor 1301 and optionally connected via storage control unit 1307. Fixed storage 1308 provides additional data storage capacity and may also include any suitable tangible computer-readable medium described herein. Storage 1308 may be used to store operating system 1309, executable file 1310, data 1311, application 1312, etc. Storage 1308 may also include optical disc drive, solid-state storage device (e.g., flash memory-based system), or any combination thereof. Where appropriate, information in storage 1308 may be incorporated into memory 1303 as virtual memory.

[0083] In one example, storage device 1335 may be removably coupled to computer system 1300 via storage device interface 1325 (e.g., via an external port connector (not shown)). Specifically, storage device 1335 and associated machine-readable medium may provide non-volatile and / or volatile storage for machine-readable instructions, data structures, program modules, and / or other data for computer system 1300. In one example, software may reside wholly or partially within the machine-readable medium on storage device 1335. In another example, software may reside wholly or partially within processor 1301.

[0084] Bus 1340 connects multiple subsystems. In this document, a bus as referred to may include one or more digital signal lines that perform a common function where appropriate. Bus 1340 can be any of several types of bus architectures, including but not limited to memory buses, memory controllers, peripheral buses, local buses, and any combinations thereof using any of a variety of bus architectures. By way of example and not limitation, such architectures include Industry Standard Architecture (ISA) buses, Enhanced ISA (EISA) buses, Micro Channel Architecture (MCA) buses, Video Electronics Standards Association Local Bus (VLB), Peripheral Component Interconnect (PCI) buses, PCI-Express (PCI-X) buses, Accelerated Graphics Port (AGP) buses, HyperTransport (HTX) buses, Serial Advanced Technology Attachment (SATA) buses, and any combinations thereof.

[0085] Computer system 1300 may also include input device 1333. In one example, a user of computer system 1300 may input commands and / or other information into computer system 1300 via input device 1333. Examples of input device 1333 include, but are not limited to, alphanumeric input devices (e.g., keyboard), pointing devices (e.g., mouse or touchpad), touchpads, touchscreens, multi-touch screens, joysticks, styluses, game controllers, audio input devices (e.g., microphones, voice response systems, etc.), optical scanners, video or still image capture devices (e.g., cameras), and any combination thereof. In some embodiments, the input device is Kinect, Leap Motion, etc. Input device 1333 may engage with bus 1340 via any of a variety of input interfaces 1323 (e.g., input interface 1323), including but not limited to serial, parallel, game port, USB, Firewire, Thunderbolt, or any combination thereof.

[0086] In a particular implementation, when computer system 1300 is connected to network 1330, computer system 1300 can communicate with other devices connected to network 1330, particularly with mobile devices and enterprise systems, distributed computing systems, cloud storage systems, cloud computing systems, etc. Communication to and from computer system 1300 can be sent through network interface 1320. For example, network interface 1320 can receive incoming communication (such as requests or responses from other devices) from network 1330 in the form of one or more packets (such as Internet Protocol (IP) packets), and computer system 1300 can store the incoming communication in memory 1303 for processing. Similarly, computer system 1300 can store outgoing communication (such as requests or responses to other devices) in memory 1303 in the form of one or more packets and communicate from network interface 1320 to network 1330. Processor 1301 can access these communication packets stored in memory 1303 for processing.

[0087] Examples of network interface 1320 include, but are not limited to, network interface cards, modems, and any combination thereof. Examples of network 1330 or network segment 1330 include, but are not limited to, distributed computing systems, cloud computing systems, wide area networks (WANs) (e.g., the Internet, corporate networks), local area networks (LANs) (e.g., networks associated with offices, buildings, campuses, or other relatively small geographical areas), telephone networks, direct connections between two computing devices, peer-to-peer networks, and any combination thereof. Networks such as network 1330 can employ wired and / or wireless communication modes. Generally, any network topology can be used.

[0088] Information and data can be displayed via display 1332. Examples of display 1332 include, but are not limited to, cathode ray tube (CRT), liquid crystal display (LCD), thin-film transistor liquid crystal display (TFT-LCD), organic liquid crystal display (OLED), such as passive matrix OLED (PMOLED) or active matrix OLED (AMOLED) displays, plasma displays, and any combination thereof. Display 1332 can be coupled to processor 1301, memory 1303 and fixed storage 1308, and other devices such as input device 1333 via bus 1340. Display 1332 is linked to bus 1340 via video interface 1322, and data transmission between display 1332 and bus 1340 can be controlled via graphical control 1321. In some embodiments, the display is a video projector. In some embodiments, the display is a head-mounted display (HMD), such as a VR headset. In further embodiments, as non-limiting examples, suitable VR headsets include HTC Vive, Oculus Rift, Samsung Gear VR, Microsoft HoloLens, Razer OSVR, FOVE VR, Zeiss VR One, AvegantGlyph, Freefly VR headsets, etc. In further embodiments, the display is a combination of devices such as those disclosed herein.

[0089] In addition to the display 1332, the computer system 1300 may include one or more other peripheral output devices 1334, including but not limited to audio speakers, printers, storage devices, and any combination thereof. Such peripheral output devices may be connected to the bus 1340 via an output interface 1324. Examples of output interfaces 1324 include, but are not limited to, serial ports, parallel connections, USB ports, firewire ports, thunderbolt ports, and any combination thereof.

[0090] Additionally or alternatively, computer system 1300 may provide functionality as a result of logic hardwired or otherwise embodied in circuitry—which may replace or operate in conjunction with software to perform one or more processes or steps of one or more processes described or illustrated herein. Software referred to in this disclosure may include logic, and the logic referred to may include software. Additionally, where appropriate, referenced computer-readable media may include circuitry (such as an IC) storing software for execution, circuitry embodying logic for execution, or both. This disclosure encompasses any suitable combination of hardware, software, or both.

[0091] Those skilled in the art will understand that the various illustrative logic blocks, modules, circuits, and algorithm steps described in conjunction with the embodiments disclosed herein can be implemented as electronic hardware, computer software, or a combination of both. To clearly illustrate this interchangeability between hardware and software, the various illustrative components, blocks, modules, circuits, and steps have been generally described above in terms of their functionality.

[0092] The various illustrative logic blocks, modules, and circuits described in conjunction with the embodiments disclosed herein can be implemented or executed using a general-purpose processor, digital signal processor (DSP), application-specific integrated circuit (ASIC), field-programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. The general-purpose processor may be a microprocessor, but alternatively, the processor may be any conventional processor, controller, microcontroller, or state machine. The processor may also be implemented as a combination of computing devices, such as a combination of a DSP and a microprocessor, multiple microprocessors, one or more microprocessors combined with a DSP core, or any other such configuration.

[0093] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be embodied directly in hardware, in a software module executed by one or more processors, or in a combination of both. The software module can reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art. An exemplary storage medium is coupled to a processor such that the processor can read information from and write information to the storage medium. Alternatively, the storage medium can be integrated into the processor. The processor and storage medium can reside in an ASIC. The ASIC can reside in a user terminal. Alternatively, the processor and storage medium can reside as discrete components in the user terminal.

[0094] Based on the description herein, suitable computing devices, by way of non-limiting example, include server computers, desktop computers, laptop computers, notebook computers, mini-notebook computers, netbook computers, internet-enabled tablet computers, set-top box computers, media streaming devices, handheld computers, internet-connected appliances, mobile smartphones, tablet computers, personal digital assistants, video game consoles, and vehicles. Those skilled in the art will also recognize that selected televisions, video players, and digital music players with optional computer networking connectivity are suitable for use in the systems described herein. In various embodiments, suitable tablet computers include those computers known to those skilled in the art that have platters, boards, and convertible configurations.

[0095] In some implementations, the computing device includes an operating system configured to execute executable instructions. For example, an operating system is software, including programs and data, that manages the device's hardware and provides services for the execution of applications. Those skilled in the art will recognize that, as non-limiting examples, suitable server operating systems include FreeBSD, OpenBSD, etc. Linux Mac OS X Windows and Those skilled in the art will recognize that, by way of non-limiting example, suitable personal computer operating systems include Mac OS and UNIX-like operating systems, such as In some implementations, the operating system is provided by cloud computing. Those skilled in the art will also recognize, by way of non-limiting example, suitable mobile smartphone operating systems include... OS Research In BlackBerry Windows OS Windows OS and Those skilled in the art will also recognize that, by way of non-limiting example, suitable media streaming device operating systems include Apple. Google Google Amazon and Those skilled in the art will also recognize that, by way of non-limiting example, suitable video game console operating systems include Xbox Microsoft Xbox One, Wii and

[0096] Non-transitory computer-readable storage medium

[0097] In some embodiments, the platforms, systems, media, and methods disclosed herein include one or more non-transitory computer-readable storage media encoded with programs comprising instructions executable by an operating system of an optionally networked computing device. In further embodiments, the computer-readable storage medium is a tangible component of the computing device. In still further embodiments, the computer-readable storage medium may optionally be removable from the computing device. In some embodiments, as non-limiting examples, the computer-readable storage medium includes CD-ROMs, DVDs, flash memory devices, solid-state storage, disk drives, tape drives, optical disc drives, distributed computing systems, including cloud computing systems and services, etc. In some cases, the programs and instructions are permanently, substantially permanently, semi-permanently, or non-transitory encoded on the medium.

[0098] Computer program

[0099] In some embodiments, the platforms, systems, media, and methods disclosed herein include at least one computer program or its use. A computer program includes a sequence of instructions that is executed by one or more processors of a computing device's CPU and is written to perform a specified task. Computer-readable instructions can be implemented as program modules that perform a specific task or implement a specific abstract data type, such as functions, objects, application programming interfaces (APIs), computational data structures, etc. In view of the disclosure provided herein, those skilled in the art will recognize that computer programs can be written in various versions of various languages.

[0100] The functionality of computer-readable instructions can be combined or distributed as needed in various environments. In some embodiments, a computer program includes a sequence of instructions. In some embodiments, a computer program includes multiple sequences of instructions. In some embodiments, the computer program is provided from one location. In other embodiments, the computer program is provided from multiple locations. In various embodiments, the computer program includes one or more software modules. In various embodiments, the computer program includes, in part or in whole, one or more web applications, one or more mobile applications, one or more standalone applications, one or more web browser plugins, extensions, accessory components, or add-ons, or combinations thereof.

[0101] Software Module

[0102] In some embodiments, the platforms, systems, media, and methods disclosed herein include software, server, and / or database modules, or their use. Given the disclosure provided herein, the software modules are created using techniques known to those skilled in the art, employing machines, software, and languages ​​known in the art. The software modules disclosed herein are implemented in various ways. In various embodiments, the software module includes files, code segments, programming objects, programming structures, or combinations thereof. In further embodiments, the software module includes multiple files, multiple code segments, multiple programming objects, multiple programming structures, or combinations thereof. In various embodiments, as non-limiting examples, one or more software modules include web applications, mobile applications, and standalone applications. In some embodiments, the software module is within a computer program or application. In other embodiments, the software module is within more than one computer program or application. In some embodiments, the software module resides on a single machine. In other embodiments, the software module resides on more than one machine. In further embodiments, the software module resides on a distributed computing platform such as a cloud computing platform. In some embodiments, the software module resides on one or more machines located at one location. In other embodiments, the software module resides on one or more machines located at more than one location.

[0103] database

[0104] In some embodiments, the platforms, systems, media, and methods disclosed herein include one or more databases, or their use. Given the disclosure provided herein, those skilled in the art will recognize that many databases are suitable for storing and retrieving 3D images and fluorescent droplet candidates. In various embodiments, suitable databases, as non-limiting examples, include relational databases, non-relational databases, object-oriented databases, object databases, entity-relational model databases, association databases, and XML databases. Other non-limiting examples include SQL, PostgreSQL, MySQL, Oracle, DB2, and Sybase. In some embodiments, the database is Internet-based. In further embodiments, the database is network-based. In still further embodiments, the database is cloud-based. In a particular embodiment, the database is a distributed database. In other embodiments, the database is based on one or more local computer storage devices.

[0105] Terms and Definitions

[0106] Unless otherwise defined, all technical terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. As used in this specification and the appended claims, the singular forms “a,” “an,” and “described” include plural references unless the context clearly specifies otherwise. Unless otherwise stated, any reference to “or” herein is intended to cover “and / or.”

[0107] Whenever the terms "at least," "greater than," or "greater than or equal to" precede the first value in a series of two or more values, the terms "at least," "greater than," or "greater than or equal to" apply to each value in the series. For example, greater than or equal to 1, 2, or 3 is equivalent to greater than or equal to 1, greater than or equal to 2, or greater than or equal to 3.

[0108] Whenever the terms “not greater than,” “less than,” “less than or equal to,” or “at most” precede the first value in a series of two or more values, the terms “not greater than,” “less than,” “less than or equal to,” or “at most” apply to each value in that series. For example, less than or equal to 3, 2, or 1 is equivalent to less than or equal to 3, less than or equal to 2, or less than or equal to 1.

[0109] Whenever a value is described as a range, it should be understood that such disclosure includes disclosing all possible subranges within such a range, as well as specific values ​​falling within such a range, whether or not the specific value or specific subrange is explicitly stated.

[0110] Example

[0111] The following examples are illustrative and are not limited to the scope of the devices, systems, fluid devices, kits, and methods described herein.

[0112] Example 1: Three-dimensional imaging of multiple tubes using light sheet imaging

[0113] This embodiment describes imaging multiple tubes using light-sheet imaging. In the first measurement, four PCR tubes arranged in a row, containing suspensions of fluorescent and non-fluorescent lipid droplets, are placed on a translation stage. The translation stage is positioned so that one of the four PCR tubes is held at the intersection of the light-sheet illumination path and the camera imaging path, which is oriented at a 90° angle relative to that illumination path, as shown below. Figure 1 As shown in the diagram, the translation stage is configured to translate along a single axis at a 45° angle relative to both the illumination and imaging paths. The tube is partially immersed in the open sample chamber, as shown in the diagram. Figure 2 As shown in A and as Figure 2As shown in Figure B, the sample chamber contains a refractive index-matching liquid that matches the lipid droplet suspension. The sample chamber has three flat, transparent surfaces positioned within the imaging path, such that the first transparent surface is positioned orthogonal to the imaging path (“Transparent Surface 1”), and the second transparent surface is positioned orthogonal to the light sheet illumination path (“Transparent Surface 2”), as shown in Figure B. Figure 1 As shown in the diagram. A third transparent surface (“transparent surface 3”) is positioned in the light sheet illumination path to allow the light sheet to illuminate the sample chamber. During imaging, the tubes are translated along the translation axis so that the continuous cross-sectional planes of each tube are illuminated and imaged. As the tubes are translated along the 45° translation axis, the positioning of the tube cross-sectional plane image shifts within the camera's field of view, as shown in the diagram. Figure 1 The diagrams (i), (ii), and (iii) show the cross-sectional planes (i) taken from the front edge of the tube, (ii) taken from the middle of the tube, and (iii) taken from the rear edge of the tube. The tube is further translated along a 45° translation axis to move the subsequent tube to the position to be imaged.

[0114] Example 2: Three-dimensional imaging of multiple tubes in a closed sample holder using light sheet imaging.

[0115] This embodiment describes the use of light sheet imaging to perform three-dimensional imaging of multiple tubes in a closed sample holder.

[0116] In this assay, four PCR tubes, arranged in a row, containing suspensions of fluorescent and non-fluorescent lipid droplets, are placed in the sample chamber on a translation stage, as shown below. Figure 3 As shown in Figure A, the positioning translation stage is used to hold one of the four PCR tubes at the intersection of the light sheet illumination path and the camera imaging path, which is oriented at a 90° angle relative to that illumination path, as shown in Figure A. Figure 3 As shown in B, the translation stage is configured to translate along two orthogonal axes—the y-axis parallel to the illumination path and the z-axis parallel to the imaging axis. Figure 3As shown in Figure A, the tube is partially immersed in a sample chamber containing a refractive index-matching fluid that matches the lipid droplet suspension and fitted with a cap with an O-ring seal. The sample chamber has two flat, transparent surfaces for each tube, plus an additional flat, transparent surface. The sample chamber is positioned in the imaging path such that a first transparent surface (“first transparent surface”) corresponding to the first sample tube is positioned perpendicular to the imaging path, and a second transparent surface (“second transparent surface”) corresponding to the first sample tube is positioned perpendicular to the light sheet illumination path. The additional transparent surface (“additional transparent surface”) is positioned in the light sheet illumination path to allow light sheet illumination to exit the sample chamber. During imaging, the tube and sample chamber are translated along the z-axis such that a continuous cross-sectional plane of each tube is illuminated and imaged. The position of the tube cross-sectional plane image remains centered within the camera's field of view during translation, compared to the first measurement. Once a tube has been scanned, the tube and sample chamber are translated along the y-axis to move a new tube to the position to be imaged.

[0117] In an alternative configuration, the tube and sample chamber are translated at 45° angles to the illumination and imaging paths along the y and z axes, such that a continuous cross-sectional plane of each tube is illuminated and imaged, as described in Example 1. Once a tube has been scanned, the tube and sample chamber are further translated along the y and z axes to move a new tube to the position to be imaged.

[0118] Example 3

[0119] Automated detection and quantification of fluorescently labeled droplets using light-sheet imaging

[0120] This embodiment describes the automated detection and quantification of fluorescently labeled droplets imaged using light sheet imaging. As described in Embodiment 1 and as... Figures 5A-5G As shown, samples containing fluorescently labeled droplets are imaged. The resulting cross-sectional planar images are combined to generate a three-dimensional image dataset for each sample. Regions of interest within each cross-section are selected to delineate the sample boundaries. Figure 5A This is to ensure that only fluorescent particles within the tube cross-section are considered. Local background subtraction is performed to suppress background fluorescence. Figure 5B A Gaussian filter is applied to smooth the image, and a medium filter is applied to remove hot pixels.

[0121] Three-dimensional (3D) convolution is performed using a spherical template to identify substantially spherical fluorescent particles (corresponding to fluorescein droplets) within a desired size range and to eliminate fluorescent particles that do not conform to the desired size and shape parameters. Figure 5C 3D local maxima analysis was performed to identify potential fluorescent particle candidates with positive fluorescence signals. Figure 5D Perform additional 3D local maxima analysis to filter out asymmetric particle candidates (e.g., non-spherical particle candidates).

[0122] Define the boundary between the bulk sample and the top layer of the sample (called the "blurred layer"), which has a lower signal than the bulk sample due to refractive index mismatch at the air-sample interface. Figure 5E The region above the blur layer boundary is grouped into a grid for separate local histogram analysis from the grid of the image containing large samples. Each tube section is divided into multiple grids for individual thresholding. Figure 5F For each grid cell, plot the logarithmic sum of the intensities of each fluorescent particle candidate. 10 (log 10 Histogram of intensity. Automatically set the threshold log for each grid cell. 10 (Intensity) values ​​are used to distinguish noise from signal. A threshold is set based on the following assumptions: (A) most particles belong to a population with some outliers (e.g., a population of lipid droplets that are mostly negative for the signal, or a population of lipid droplets that are mostly positive for the signal), or (B) two populations of particles exist. For case (A), a median plus or minus standard deviation threshold is used to automatically separate positive and negative signal particles. For case (B), an optimization algorithm is used to select a threshold that maximizes the separation between the positive signal population and the threshold, such as by measuring the number of standard deviations of the positive population above the threshold.

[0123] The ratio of positive to negative signal particles is determined for each grid cell. It is expected that the ratio of positive to negative signal particles should not vary significantly between grid cells, despite differences in brightness. A large deviation of the ratio in a given grid cell from the median ratio indicates an inappropriate threshold selection. Grid cells with ratios significantly higher or lower than the median ratio are identified as outliers and are automatically selected for an additional round of refinement and threshold detection. Figure 5G After further refinement, the remaining outliers are removed, and the global positive particle count is determined.

[0124] Example 4

[0125] High-throughput digital PCR system

[0126] This embodiment describes a light-sheet imaging setup and a method for automatically imaging eight or more digital PCR samples without user intervention using such a setup. The setup described herein is designed to allow imaging of eight PCR samples using eight sample strips. However, it should be noted that this setup can be reconfigured to allow analysis of fewer or more samples at a time.

[0127] Figure 6 The procedure settings and steps for sample loading and scanning for this high-throughput digital PCR analysis method are shown. Figure 6As shown in A, 96 PCR samples are loaded onto a sample loading block using, for example, a 96-well plate (as indicated by arrow #1), and samples can be picked from the sample loading block in 8-sample strips, thus picking 8 samples at a time. Next, as... Figure 6 As shown in B, eight sample strips (as indicated by arrow #2) are picked up from the sample plate and moved on a translation stage (as indicated by arrow #3) to scan the samples. The PCR samples are then scanned using a focusing sheet from a collimated laser source (as indicated by white arrow #4) to generate a light sheet image of each of the eight samples on the sample strips. Using a resonant mirror, the light sheet is configured to pivot and / or translate at a frequency of approximately 1 kHz while scanning the samples to improve image quality by reducing artifacts, as described in Example 5 below. After scanning, the eight sample strips are moved back to the sample loading block, and a second eight sample strip is picked up and imaged.

[0128] Example 5

[0129] Improving image quality by reducing artifacts

[0130] This embodiment describes a method for improving image quality in light sheet imaging experiments. Light scattering and absorption during scanning can cause banding and shadow artifacts in light sheet images. To reduce these artifacts, the imaging setup is configured (e.g., by using resonant mirrors) such that the light source (e.g., a laser light sheet) can make some movement relative to the sample during scanning, thereby eliminating or "averaging out" some (or most) of the scattering and absorption events. Such movement of the light source relative to the sample (e.g., a digital PCR sample) is in the kHz range (e.g., from about 0.1 kHz to about 20 kHz) and as... Figure 7 As shown in the image. Figure 7 Figure A shows how pivoting the light plate improves image quality, such as... Figure 7 As shown in the rightmost "Average" image. Figure 7 B shows that shifting the light plate improves image quality, such as... Figure 7 The rightmost "average" image is shown in B.

[0131] Another way to reduce artifacts is to use double-sided lighting, such as... Figure 7 As shown in C. Here, the sample (e.g., a PCR sample tube) is illuminated from both sides, such that each light source only needs to pass through half of the sample to illuminate the entire sample cross-section.

[0132] Example 6

[0133] Improving image quality by reducing artifacts

[0134] This embodiment describes the automated detection and quantification of fluorescently labeled droplets imaged using light sheet imaging. As described in Embodiment 1 and as... Figures 5A-5GAs shown, samples containing fluorescently labeled droplets are imaged. The resulting cross-sectional planar images are combined to generate a three-dimensional image dataset for each sample. Regions of interest within each cross-section are selected to delineate the sample boundaries. Figure 5A This is to ensure that only fluorescent particles within the tube cross-section are considered. Local background subtraction is performed to suppress background fluorescence. Figure 5B A Gaussian filter is applied to smooth the image, and a medium filter is applied to remove hot pixels.

[0135] Three-dimensional (3D) convolution is performed using a spherical template to identify substantially spherical fluorescent particles (corresponding to fluorescein droplets) within a desired size range and to eliminate fluorescent particles that do not conform to the desired size and shape parameters. Figure 5C 3D local maxima analysis was performed to identify potential fluorescent particle candidates with positive fluorescence signals. Figure 5D Perform additional 3D local maxima analysis to filter out asymmetric particle candidates (e.g., non-spherical particle candidates).

[0136] Candidate fluorescent droplets were determined based on 3D local maxima analysis. Figure 8A The convolutional neural network (CNN) is used to distinguish noise from signal by determining the shape and size data of the fluorescent droplets as a whole. Figure 8B Clusters are then removed to eliminate overlapping fluorescent droplets with peak densities below a set threshold. Peak density is determined based on droplet size, which is calculated based on the full width at half maximum (WHM). Cluster removal is based on nearest neighbor / volume exclusion / non-maximum suppression and by determining the number of close droplet neighbors for each droplet. A histogram of close droplet neighbors is generated. Figure 8C ), and eliminate abnormal droplets with a certain number of neighbors exceeding a set threshold.

[0137] While preferred embodiments of the present disclosure have been shown and described herein, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many variations, modifications, and substitutions will now occur to those skilled in the art without departing from the present disclosure. It should be understood that various alternatives to the embodiments of the present disclosure described herein may be employed in carrying out the present disclosure. The appended claims are intended to define the scope of the present disclosure and thereby cover the methods and structures within the scope of these claims and their equivalents.

Claims

1. A light-sheet imaging system, comprising: Multiple containers containing multiple samples, wherein each of the multiple samples contains a distribution of fluorescent particles and is contained within a container of the multiple containers; A collimated illumination source, configured to emit a collimated beam along the beam path; A Powell lens, positioned on the beam path after the collimating illumination source, and configured to extend the collimated beam along a single axis; A first cylindrical lens is positioned on the beam path behind the Powell lens and configured to recollimate the beam along a single axis, thereby producing a collimated beam having a short axis orthogonal to the propagation direction and a long axis orthogonal to both the short axis and the propagation direction. as well as A second cylindrical lens is positioned on the beam path behind the first cylindrical lens and configured to focus the beam along the minor axis onto the focal plane. The light sheet imaging system is configured to automatically scan a set of cross sections of each of the plurality of samples within each of the plurality of containers using a camera.

2. The light sheet imaging system of claim 1, wherein the camera includes an imaging path orthogonal to the beam path, wherein the camera is positioned such that the imaging path intersects the beam path at the focal plane.

3. The light sheet imaging system according to claim 1, wherein the plurality of containers comprises a plurality of tubes.

4. The light sheet imaging system of claim 3 further includes a tube holder configured to hold the plurality of tubes.

5. The light sheet imaging system of claim 4 further includes a fluid reservoir configured to contain fluid.

6. The optical sheet imaging system of claim 5, wherein the fluid reservoir and the tube frame are partially immersed in the fluid.

7. The light-film imaging system of claim 5 further includes a cover configured to be mounted on top of the fluid reservoir.

8. The light sheet imaging system of claim 5, wherein the fluid reservoir includes a first transparent surface and a second transparent surface connected to the first transparent surface.

9. The light-film imaging system according to claim 3, wherein the tubes in the plurality of tubes comprise PCR tubes or microcentrifuge tubes.

Images