Radiant energy – Photocells; circuits and apparatus – Optical or pre-photocell system
Reexamination Certificate
2000-03-16
2002-06-11
Allen, Stephone (Department: 2878)
Radiant energy
Photocells; circuits and apparatus
Optical or pre-photocell system
C356S417000
Reexamination Certificate
active
06403947
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to equipment and methods for imaging a sample which may contain one or more species of luminous probes such as fluorescent probes, luminescent probes, quantum dot probes, up-converting probes, or other emissive probes; and more specifically, to equipment and methods which can image N species of probes with high optical efficiency, often requiring fewer than N observations of the sample.
2. Description of the Related Art
It is often of practical importance to image multiple luminous probes in a single sample, as in fluorescence in-situ hybridization (FISH), in chromatography plate readers, DNA sequencing, spectral karyotyping, general biological and neurobiological research, and the like.
One approach to multiprobe imaging, taken by Glass et. al. in U.S. Pat. No. 5,723,294 is simply to image the sample, in this case a multiwell plate, in several plate readers one after another. By suitable choice of probes and reader settings, unambiguous readings can be obtained from the combined results of all the readers. Yet this is hardly an integrated, efficient solution to the problem.
There is also a large literature of hyperspectral imaging, which can be used to obtain multiprobe images of samples. In hyperspectral imaging, several exposures of a sample are recorded using filters or interferometers in the optical path, from which an optical spectrum is derived for each point in the sample under study. Specific hardware for hyperspectral imaging includes filter wheels and circular-variable filters as in U.S. Pat. No. 5,591,981 and U.S. Pat. No. 5,784,152; angle-tuned interference filters as in the Renishaw imaging Raman microscope described in U.S. Pat. No. 5,442,438; acousto-optical tunable filters (AOTFs) as in U.S. Pat. No. 5,216,484, U.S. Pat. No. 5,377,003, and U.S. Pat. No. 5,556,790; optical interferometers as in U.S. Pat. No. 5,835,214, U.S. Pat. No. 5,817,462, U.S. Pat. No. 5,539,517, and U.S. Pat. No. 5,784,162; and liquid crystal tunable filters (LCTFs) as in Morris, et. al., “Imaging Spectrometers for Fluorescence and Raman Microscopy: Acousto-Optic and Liquid Crystal Tunable Filters,” Applied Spectroscopy, 48:7:857-866, 1994. All of these except the interferometer systems are termed band-sequential systems, as each image records the entire spatial content of the sample, and successive images serve to step through its spectral content.
Alternatively, dispersive systems are used to obtain a spectrum for a single point or line, which is then scanned in two or one dimension respectively, to obtain a 2-D image of the sample with spectral data for each point. A non-dispersive system is described by Buican et. al., who use a photoelastic modulator (PEM) and polarizer in U.S. Pat. No. 4,905,169 to determine the spectral contents of a single point via the Fourier analysis of time-series intensity values at a detector; in U.S. Pat. No. 5,117,466 this arrangement is coupled with a laser scanning system to produce a two-dimensional image. Such systems are termed point-sequential or line-sequential imagers, as the entire spectral content is recorded more or less simultaneously, and successive readings step through the spatial content either pointwise or a line at a time.
Prior art describes fluorescence imaging where the excitation and/or emission selection is set to discrete wavelength settings (U.S. Pat. No. 5,784,152), and where the wavelength selection is continuously tunable (U.S. Pat. No. 5,591,981 and U.S. Pat. No. 5,863,504). Excitation light tuning is achieved by filter means such filter wheels, AOTF's, LCTF's; or via galvanometer-driven gratings; or via a series of paddle-mounted filters; or two arc lamps, each of which has control means for rapidly adjusting its intensity over a wide range (U.S. Pat. No. 5,491,343). In U.S. Pat. No. 5,208,651, Buican describes a method for time-encoding the excitation spectrum while concurrently analyzing the emission spectrum, via two PEM elements.
Normally, the individual bands used in hyperspectral imaging are distinct or nearly so, overlapping only in the transition region where a given band cuts off and the adjacent band cuts on. Each band has a transmission vs. wavelength response that approximates a steep-edged trapezoid, with sharp cut-on and cut-off, and approximately constant transmission through the passband. It is a universal goal in hyperspectral imaging to maximize the transmission at all wavelengths in the passband, as this yields increased signal-to-noise, which is a general concern in the field. Researchers have developed methods for broadening the inherently narrow bandpass of the AOTF, to obtain an approximately trapezoidal bandpass instead of a narrow sync function.
The prior art includes other methods for increasing throughput to obtain better signal-to-noise, such as the use of LCTF filters based on Solc designs, with broad bandpasses for increased throughput (Hoyt, “Tunable Liquid Crystal Filters Boost Fluorescence Imaging”, BioPhotonics, July/August 1996). This may result in some overlap, or crosstalk, between adjacent spectral bands, which is dealt with by methods such as those described in the next three paragraphs. Notwithstanding the overlap between bands, this art is practiced with trapezoidal bandpass shapes or the like, having the highest practical transmission in each passband.
Many integrated multiprobe readers seek to take advantage of a priori knowledge of the samples being imaged. As the probes have more or less predetermined spectra, a complete spectrum may not be required. For example, it is not necessary to produce spectral data for those wavelength bands at which there is no possibility of optical emission. At the same time, the emission spectra of the various probes involved are not always distinct, but may also overlap to a considerable degree in some cases. If it is not possible to choose a set of wavelengths that correspond in a one-to-one fashion with the probes being imaged, the presence of emission at any given wavelength does not uniquely specify which probe was present. Rather, for a given experimental set-up, the observed energy e
i
at wavelength band &lgr;
i
is related to the concentration of the various probes c
j
according to:
e
i
=a
i1
*c
1
+a
i2
*c
2
+a
i3
*c
3
. . . a
iN
*c
N
[1]
where coefficient a
ij
specifies the optical radiation of probe j into wavelength band i.
This provides an easy way to determine the probe concentrations from the observed intensities, as follows. Equation [1] may be written in matrix form:
E=A*C
[2]
where E is the M×1 vector of observed energies at the M spectral bands, A is the M×N matrix of terms a
ij
, and C is the 1×N vector of probe concentrations. The matrix A has a direct physical interpretation. Each column corresponds to the spectrum of each particular probe, while each row corresponds to the emission of the various probes at a particular wavelength band. While some systems use a number of wavelength bands M greater than the number of probes N, it is common to use M=N, and to seek wavelengths where this matrix is approximately diagonal. Physically this means that for every probe, there is a corresponding wavelength band for which most, though not all, of the optical energy comes from that probe.
Since the number of wavelength bands M equals or exceeds the number of probes N, equation [2] may be inverted uniquely or in a least-squares error fashion to solve for the probe concentrations, viz.:
C=A
−1
*E
[3]
This allows direct calculation of the probe concentrations from the observed spectra, despite the presence of overlap or cross-talk between the spectral bands. Using this approach, Castleman used 3×3 matrices to produce images of three fluorescent probes from three raw intensity images of sample emission, in “Color Compensation for Digitized FISH images,” Bioimaging, 1:159-165, 1993, and again in “Digi
Hoyt Clifford C.
Levenson Richard M.
Miller Peter J.
Cambridge Research & Instrumentation Inc.
Cohen & Pontani, Lieberman & Pavane
LandOfFree
High-efficiency multiple probe imaging system does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with High-efficiency multiple probe imaging system, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and High-efficiency multiple probe imaging system will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2931834