METHOD AND APPARATUS FOR REMOVAL OF NON- UNIFORMITIES IN AN ELECTROPHORESIS APPARATUS
BACKGROUND OF THE INVENTION The present invention relates generally to electrophoresis reading systems and, more particularly, to a method and apparatus for removing non-uniformities in an electrophoresis system. In the biotechnical field, fluorescent dyes are routinely used as sensitive, non-isotopic labels. These labels are used to identify and locate a variety of cell structures, ranging from malignant tumors to specific chromosomes in a DNA sequence. A variety of devices have been designed to read fluorescent-labeled samples.
Gel electrophoresis is one technique commonly used in conjunction with fluorescent dyes and other markers to identify specific molecules as well as other tagged units. In this technique an electric field is used to cause the migration of the tagged units through a gel or other solution.
In U.S. Patent No. 4,874,492 a gel electrophoresis system is disclosed in which samples are treated with fluorescent markers prior to applying them to an electrophoretic gel. The gel is illuminated with a UV source and the fluorescence pattern is detected with a cooled charge-coupled-device (CCD) two-dimensional detector array. The CCD array is cooled to at least -25 degrees C. in order to improve light sensitivity and increase the dynamic range.
In U.S. Patent No. 5,162,654 a system is disclosed to optically determine which of four fluorophores is fluorescing in an electrophoresis gel. Fluorescence emitted by the gel passes first through four separate band pass filters and then through four wedge prisms. As a result of this optical configuration, the emitted fluorescence is imaged on four discrete areas on the detector array. The specific fluorophore exited by the irradiation source is determined by comparing the relative intensities of the fluorescence detected in the four detection areas.
In U.S. Patent No. 5,294,323 the disclosed gel electrophoresis system utilizes a vertical electrophoresis plate. A laser beam passes horizontally through the gel in a direction perpendicular to the longitudinal axis of the electrophoresis plate. The emitted fluorescence is reflected to a solid state imaging sensor such that the reflected pattern is parallel to the direction of the laser beam.
In U.S. Patent No. 5,324,401 a fluorescence detection system for capillary electrophoresis is disclosed which provides for the simultaneous excitation and detection of fluorescent probes within a plurality of capillaries. The excitation source is a laser which is coupled to the capillaries through an optical fiber bundle. The fluorescence from the capillary array is focussed through a lens and imaged onto a CCD camera for analysis. In a paper by Sutherland et al. entitled "Electronic Imaging System for Direct and Rapid Quantitation of Fluorescence from Electrophoretic Gels: Application to Ethidium Bromide-Stained DNA" published in Analytical Biochemistry 163, 446-457 (1987), the authors describe an imaging system which uses a CCD camera. The CCD camera quantifies the fluorescence received from electrophoretic gels, chromatograms, and other sources. The paper describes several sources of non-uniformities which impact the ability of the system to obtain accurate results.
From the foregoing, it is apparent that an improved electrophoresis apparatus is desired which enables accurate quantitative fluorescence measurements.
SUMMARY OF THE INVENTION The present invention provides a method and apparatus for correcting for the non-uniformities in an electrophoresis apparatus. These non-uniformities arise from the imaging system. By correcting for these non-uniformities it is possible to make quantitative measurements of an electrophoresis gel, thus increasing the information which can be obtained from the electrophoretic analysis.
In one aspect of the invention, the system uses a look-up table to correct a sample image of any non-uniformities arising from the lens assembly. The look-up table contains the uniformity profiles for the lens assembly for a range of aperture and magnification settings. In order to correct a sample image, the aperture and magnification settings used to obtain the sample image are provided to a system processor. These settings may be automatically obtained by the processor or manually input by the user. After the processor receives the lens settings, the look-up table is used to determine the
corresponding lens non-uniformities. Preferably the look-up table contains a correction file for each pair of lens settings. The sample image is normalized by dividing the sample image file by the appropriate correction file. Once normalized, the corrected sample image file may either be displayed or stored for later use. In another aspect of the invention, the non-uniformities due to the illumination source are characterized by sampling a portion of the source with a linear array detector. In order to sample the source, a mirror or beamsplitter is appropriately positioned, for example along the central optical axis of the electrophoresis apparatus. Preferably the mirror is brought into the sampling position with a stage whenever the source non-uniformities are to be measured. The reflected light is enlarged before it is measured with a linear detector.
Once a profile of the source has been determined, a correction file is stored within the memory of an associated data processor. Since the correction file must contain the same number of data points as the actual sample image file, a number of the correction data points must be filled-in by the system processor. These additional correction data points are determined through interpolation and by utilizing the source symmetry. After a completed correction file has been stored, an image of the sample is taken and a corresponding image file stored within memory. To create a corrected image file, the initial sample image file is normalized with the correction file. In another aspect of the invention, the non-uniformities due to the lens assembly are characterized by passing a secondary source of known uniformity through the lens assembly. If desired, the same mirror used to determine the source non- uniformities can be used to determine the lens non-uniformities. The secondary source is a linear source, either a continuous source or a series of point sources. As before, a correction file is created and stored which is used to normalize the image file.
In another aspect of the invention, the non-uniformities in the lens and detector assemblies are determined through the use of a reference or calibration standard which exhibits uniform fluorescence. An image of the standard is taken using the same lens settings as used with the unknown, thus insuring that the calibration image correctly reflects the system non-uniformities as applied to the unknown. The image of the calibration standard along with a darkfield image are used to correct the sample image. In one embodiment a calibration standard is contained within a filter wheel, the filter wheel proximate the entrance to the lens assembly. After a sample image is taken, the filter
wheel rotates in order to place the calibration standard immediately in front of the lens. An image of the calibration standard is taken followed by a darkfield image. The darkfield image is taken by closing a shutter to the detector, thus preventing source light, fluorescence, or any other source of illumination from reaching the detector. The image is corrected by applying a simple algorithm which uses both the calibration standard image and the darkfield image.
A further understanding of the nature and advantages of the present invention may be realized by reference to the remaining portions of the specification and the drawings.
BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is an illustration of a cross-section of a gel electrophoresis apparatus according to the prior art;
Fig. 2 is the intensity profile of a single light bulb measured perpendicular to the axis of the bulb;
Fig. 3 is an illustration of an intensity profile for a source with multiple bulbs;
Fig. 4 is an illustration of a method of improving the uniformity of an illumination source; Fig. 5 is a block diagram outlining the principal steps in correcting an image for lens non-uniformities based on the present invention;
Fig. 6 is an illustration of an embodiment of the invention; Fig. 7 illustrates the relationship between the lens aperture and the size of the calibration standard; Fig. 8 illustrates the relationship between light uniformity and the calibration standard;
Fig. 9 is an illustration of an alternate embodiment of the invention utilizing a diffuser in conjunction with the calibration standard;
Fig. 10 is a detailed illustration of an embodiment of the invention; Fig. 11 is a close-up view of the detector assembly, lens assembly, and filter wheel of the embodiment illustrated in Fig. 10;
Fig. 12 is a block diagram illustrating the process steps of the invention;
Fig. 13 is an illustration of the major components of a system according to the preferred embodiment of the invention;
Fig. 14 is an illustration of another method of determining the non- uniformities of the lens assembly; Fig. 15 is an illustration of another method of determining the non- uniformities of the lens assembly;
Fig. 16 is an illustration of the major components of a system according to the present invention;
Fig. 17 is an illustration of a method used to determine the aperture and magnification settings of a lens assembly;
Fig. 18 is an illustration of an embodiment of the invention; Fig. 19 is a block diagram outlining the principal steps used in correcting an image for illumination non-uniformities in accordance with the present invention;
Fig. 20 illustrates the interpolation process used in preparing a correction file;
Fig. 21 is an illustration of the major components of a system according to the preferred embodiment of the invention;
Fig. 22 is an illustration of an alternate embodiment of the invention which can be used to measure the non-uniformities of both the illumination source and the lens assembly;
Fig. 23 is an illustration of a system according to the present invention which includes the components necessary to perform the lens non-uniformity measurements; and
Fig. 24 is a block diagram outlining the principal steps used in accordance with another embodiment of the invention.
DESCRIPTION OF THE SPECIFIC EMBODIMENTS Fig. 1 is an illustration of a cross-section of a gel electrophoresis apparatus according to the prior art. In this system a gel plate 101 is illuminated by a light source 103. Light source 103 is comprised of a plurality of individual light bulbs 105. The light from source 103 causes fluorophores or other fluorescing material contained within specific areas of sample 101 to fluoresce. The emitted fluorescence passes through one or
more lenses 107 and is imaged onto a detector 109. Based upon the received image it is possible to determine the areas of fluorescence on sample 101.
Although the intensity of the fluorescence from sample 101 contains additional information such as the quantity of the fluorescing material, to date the ability to quantify this information has been limited due to non-uniformities in the illumination source, the imaging optics, and the detector.
The detector non-uniformities are most easily removed. For example, many CCD arrays are available which offer a linear response over a wide range of intensity levels. Typically these arrays also offer a very uniform response from pixel to pixel, thus dramatically reducing any non-uniformities arising from the detector.
The light intensity from an individual light bulb 105 is relatively uniform along the majority of the length of the bulb. At either end of the bulb the brightness level exhibits a minor fall-off in intensity. This fall-off can be minimized through the use of reflectors, masks, diffusion filters, or some combination thereof. The effects of fall-off can also be minimized by simply using longer bulbs. By extending the bulbs, the end portions of the bulbs exhibiting the lower brightness levels are located past the sampling area of the apparatus, thus placing only the relatively uniform length of the bulb under the sample.
Fig. 2 is the intensity profile of a single light bulb 105 measured perpendicular to the axis of the bulb. As expected, the profile exhibits a peak 201 centered directly above the bulb with a rapid fall off in the intensity as the distance from the bulb is increased. A dip 203 is a result of a scribe mark on the stage. Fig. 3 is an illustration of an intensity profile for source 103 measured perpendicular to the axes of bulbs 105. Peaks 301 are located over the center lines of the individual bulbs 105 while valleys 303 represent the mid-points between bulbs. Further improvement in source uniformity can be achieved by increasing the number of bulbs and decreasing the separation between bulbs.
Fig. 4 is an illustration of another method of improving the uniformity of the illumination source. In this method sample 101 is illuminated with a single source 401. Although source 401, as shown, is comprised of a single bulb, source 401 may also be comprised of multiple bulbs. Source 401 is scanned in a direction 403 perpendicular to the axis of the bulb, thus taking advantage of the high degree of intensity uniformity along the central bulb portion. The effects of intensity fall-off near either end of the bulb
may be minimized by extending the bulb past the edges of sample 101 by a distance 405. Since source 401 is scanned in a direction perpendicular to the region of bulb uniformity, sample 101 will be uniformly illuminated as long as the scanning rate remains constant. Fig. 5 is a block diagram outlining the principal steps in correcting an image for lens non-uniformities based on the present invention. First, the non- uniformities of the lens assembly are measured (step 501). A number of techniques can be used to determine lens non-uniformities, a few of which are described below. After the lens assembly has been characterized for an initial set of aperture and magnification settings, these settings are changed (steps 503 and 505) and the lens is recharacterized. This process continues until the lens assembly has been characterized for a range of aperture settings and a range of magnification settings. This information is then used to create a look-up table or matrix (step 507).
After the lens characterization look-up table has been created, this data is converted to a correction file look-up table (step 509). This step may be required in order to create a one to one correspondence between the number of correction data points per correction file and the number of data points resulting from the sample imaging process. Depending upon the technique used to characterize the lens assembly, the number of correction data points may be substantially less than the number of sample points. In this case the additional correction data points are filled-in by the data processor using well known interpolation techniques and taking advantage of any symmetry of the lens assembly.
After a look-up table of correction files has been created, the system is used to image an electrophoresis sample gel (step 511). At this time the user may either view the uncorrected sample image or store a sample file of the uncorrected image for correction and later viewing. Prior to correcting the sample image, the aperture and magnification settings used to obtain the sample image must first be determined (step 513) since this data is required in order to locate the appropriate correction file within the correction file look-up table. This information can either be determined automatically by the system or the user can simply input the lens settings. To remove the lens assembly non-uniformities from the sample image, the sample file is normalized by dividing it by the appropriate correction file (step 515). Prior to normalization, the correction files may first be smoothed using a smoothing function such as those well known in the art. The smoothing function insures that any
source of noise in the correction file is not amplified by the normalization process. Lastly, the corrected sample image is either displayed (step 517) or stored (step 519) for future use.
One method of eliminating the effects of non-uniformities in the lens and in the detector is through the use of a reference or calibration standard. During calibration, the standard replaces the sample. A suitable calibration standard is a piece of fluorescent glass or plastic which uniformly fluoresces when illuminated with a source. Preferably the standard fluoresces at the same wavelength as the labels used with the sample, thus insuring that the standard accurately reproduces the non-uniformities of the lens and the detector.
In practice, the user first takes an image represented by Iι(x,y) of the sample in question. The sample is then replaced with the standard and a second image, I2(x,y), is taken. The second image is known as a flatfield image. To obtain a corrected image a third exposure must be made in which the camera shutter is blocked completely. The third image, D0(x,y), represents the darkfield. The corrected image, CI(x,y), is represented by:
CI(x,y)-[{Iι(x,y)-D0(χ,y)}/{I2(x,y)-Do(x,y)}] * M
where M is equal to the average I2(x,y).
Although the above technique can be used to obtain accurate and therefore quantifiable images of the sample, assuming a uniform illumination source, the technique is impractical for many applications. The technique requires the user to change samples and make a series of measurements in order to obtain a single corrected image. Fig. 6 is an illustration of another embodiment of the invention. As in the prior system shown in Fig. 1, fluorescence from a sample 101 is imaged onto a detector 109 by a lens 107. In this embodiment, if the user wishes to obtain a corrected image, a calibration standard 601 is moved into place with a stage 603. In the preferred embodiment of the invention, standard 601 is placed in close proximity to the entrance aperture of lens 107. Preferably stage 603 is a rotation stage. As in the previous technique three images are required; a sample image, a calibration image, and a darkfield image. The equation to obtain the corrected image is the same as that given above.
In order to obtain an accurate mapping of the non-uniformities of the lens, calibration standard 601 must completely fill the aperture of the lens. However, since standard 601 is much closer to lens 107 than a standard placed at the sample plane as in the previous system, standard 601 may be much smaller. This relationship is illustrated in Fig. 7. The collection aperture is represented by dotted lines 701. As illustrated, standard 601 is approximately the same size as lens 107 due to its close proximity to the lens. However, a standard 703 placed at the sample plane must be much larger in order to fill the lens aperture, thus increasing fabrication costs and difficulty.
Although the present invention may be used to overcome the non- uniformities associated with the lens and the detector, in order to quantify a sample image the illumination source must be uniform. If the source is not uniform, the inaccuracies associated with the sample image, Iι(x,y), will not be completely removed by the technique of the invention. Illumination uniformity may be achieved using multiple light sources, reflectors, masks, or scanning light systems. Light uniformity is not, however, required during the calibration step of the preferred embodiment of the present invention. Although increased light uniformity may increase the accuracy of the described calibration technique, given the distance 605 separating light source 103 from standard 601, illumination uniformity is not a requirement for this technique. Fig. 8 illustrates the relationship between light uniformity and the placement of the calibration standard. A light source 801 illuminates a calibration standard 803 placed in close proximity to light source 801. As defined by Lambert's Law, the intensity of a small incremental area of the source, J0, is equivalent to the intensity of the incremental area in the direction of the normal, J0, times the cosine of the angle θ as measured from the surface normal. Therefore the intensity measured at a point 805 on standard 803 which is directly above an area 807 of source 801 is equivalent to the intensity of area 807. In contrast, the intensity measured at a point 809 at the opposite end of standard 803 is significantly reduced. For example given an angle θ of 80 degrees, the intensity at point 809 is only 17 percent of the intensity measured at point 805. Thus if source 801 is non-uniform, calibration standard 803 will be non-uniformly illuminated, resulting in a calibration technique which is a function not only of lens and detector non- uniformities, but also a function of illumination non-uniformities.
In contrast, if a calibration standard is used which is located at some distance from source 801, the illumination non-uniformities have little impact on the calibration technique. For example, assuming that a calibration standard 811 measuring 2 centimeters in diameter is located 65 centimeters from source 801, there is less than 1 percent difference in intensity as measured at points 813 and 815. Thus by placing the calibration standard near the lens and at a distance from the source, the effects of illumination non-uniformities can be greatly minimized.
Fig. 9 is an illustration of an alternate embodiment of the invention. This embodiment contains the same elements as illustrated in Fig. 6 with the addition of a diffuser 901. Diffuser 901 is mounted between calibration standard 601 and lens 107. Diffuser 901 insures that the system only monitors the non-uniformities of the lens and detector by smoothing out even the small non-uniformities within the calibration standard.
Fig. 10 is a detailed illustration of an embodiment of the invention. In this embodiment a sample 1001 is illuminated with a light source 1003. Source 1003 is comprised of three individual light bulbs 1005 which are disposed within a tray 1007. Tray 1007 rides on a pulley and belt system 1009 which is controlled by a motor 1011. During the sampling period, light tray 1007 scans sample 1001, starting at a point 1013 and ending at a point 1015. The scan rate, the intensity of the source, and any intervening filters 1017 determine the intensity of the radiation illuminating sample 1001, resulting in the fluorescence of appropriately marked regions of sample 1001. Scanning source 1003 insures that the illumination is uniform.
Emitted fluorescence from sample 1001 passes through either an aperture 1019 or a filter 1019 in a filter wheel 1021 prior to being imaged by a lens assembly 1023 onto a detector 1025. In this embodiment detector 1025 is a CCD detector, preferably a two dimensional array. When a calibration image is required, filter wheel 1021 rotates in order to place a calibration standard 1027 in front of the lens assembly.
There are several ways in which calibration standard 1027 may be illuminated. First, radiation passing through sample 1001 from source 1003 may be used. Due to the pattern on gel sample 1001, the illumination intensity will be quite non- uniform. However, as shown above, this standard is relatively insensitive to illumination non-uniformities. Source 1003 may either be used in the normal scanning mode or maintained in a stationary position at the center of the apparatus. Second, sample 1001
may be removed and source 1003 used in either a scanning or stationary mode. Third, a separate source 1029 may be used to illuminate calibration standard 1027. Preferably source 1029 is positioned above the sample plane as shown, thus not requiring the removal of sample 1001 prior to taking a calibration image. In the preferred embodiment a pair of sources 1029 are used to illuminate standard 1027, thus improving the uniformity of the light reaching standard 1027.
Fig. 11 is a close-up view of detector assembly 1025, lens assembly 1023, and filter wheel 1021. As illustrated, calibration standard 1027 has been rotated into place in readiness for a calibration measurement. Fig. 12 is a block diagram illustrating the process steps of the present invention. Initially a sample is placed in the testing apparatus (step 1201). Although the invention is not limited to gel electrophoresis applications, this is one of the most common applications. In this instance the sample is a gel in which certain regions, structures, or molecules have been labeled with a fluorescent marker which fluoresces at a specific wavelength when irradiated at the marker's excitation wavelength. The emission wavelength is different from the excitation wavelength, thus allowing the emission to be distinguished from the source through the use of filters. The emission from the selected regions is imaged onto a detector in order to form an image of the sample (step 1203). Once a sample image has been taken, a calibration standard is moved into place (step 1205). In order to properly measure the non-uniformities of the lens and detector assemblies, the system aperture and magnification settings are not changed from those used during the sample imaging step. A calibration image is taken (step 1207) using either the sample illumination source or an alternate source as described above. The camera shutter is then closed (step 1209) and a darkfield image is taken (step 1211). The sample image is then corrected (step 1213) using well known mathematical relationships such as the correction algorithm given above.
Fig. 13 is an illustration of the major components of a system according to one embodiment of the invention. In this embodiment the key system components are connected to a processor 1301. Processor 1301 permits the entire process to be automated, thus making the correction process straightforward and minimizing the risk of user errors. To use the system, sample 1001 is first placed within electrophoresis apparatus 1303. The user then manually sets the aperture and magnification of the system unless the system has been designed to automatically select these parameters. In the
preferred embodiment, the user then enters into processor 1301, through a user interface 1305, the marker (e.g., fluorophores) in use on sample 1001. Based on this information processor 1301 uses an integrated look-up table to determine the appropriate excitation and emission settings. Processor 1301 then moves an appropriate filter in place by manipulating filter wheel 1021.
Once the system operating parameters have been set, either manually or automatically, the user initiates a run by sending an appropriate signal to processor 1301 through interface 1305. Processor 1301 then turns on source 1003 and takes an image of sample 1001. If source 1003 is a scanning source as shown in Fig. 10, processor 1301 also controls the scanning operation by controlling scanning motor 1011.
The image taken by detector 1025 is stored in memory resident within processor 1301 as well as being displayed on a monitor 1307. If detector 1025 is a CCD camera, the data is easily stored in a digital format, thus making later data manipulation easy. Depending upon the configuration of processor 1301, several different operations can take place next. If the system is in a completely automatic mode, the system will automatically begin the process of correcting the sample image. In this mode an initial uncorrected image may or may not be shown, depending upon the system set-up. The benefit of showing the initial image is that the user may be able to determine that the image contains no useful data and thus end the cycle prior to the correction technique being applied.
In the automatic mode, after a sample image has been taken, processor 1301 turns off source 1003 and moves a calibration standard into place. Since the calibration standard is within filter wheel 1021, processor 1301 simply rotates the wheel until the appropriate filter recess is in place. Processor 1301 then activates the appropriate illumination source, either source 1003 or alternate sources 1029, and takes a flatfield image. The data from this image is stored in the resident processor memory. After the calibration standard image has been completed, processor 1301 closes a shutter to camera assembly 1025 and takes a darkfield image. Using these two additional images, processor 1301 corrects the sample image and presents the corrected image on monitor 1307. Both the uncorrected and the corrected images as well as the flatfield and darkfield images can be stored for later retrieval and use.
Fig. 14 is an illustration of another method of determining the non- uniformities of the lens assembly. A sample plane 1401 is illuminated with a light source
1403. In this embodiment light source 1403 consists of a series of light bulbs 1405 although this technique is equally applicable to other light sources. In normal usage a sample, for example an electrophoresis gel, would lie within sample plane 1401. The fluorescence from the excited regions of such a sample would be imaged by a lens assembly 1407 onto a detector 1409.
In order to characterize the lens assembly 1407, a secondary source 1411 is used. Source 1411 is a line source, for example a series of light emitting diodes (LEDs) 1413. Source 1411 may also consist of a series of other types of point sources or a single line source such as a long light bulb. The radiation from source 1411 is enlarged with a lens 1415 before being reflected by a mirror 1417. Mirror 1417 may be a partial reflector thus allowing it to remain in place during the normal operation of the system. Mirror 1417 may also be dedicated to the task of measuring the lens non-uniformities. If mirror 1417 is a dedicated mirror, it is placed on a stage so that it can be moved out of the sample imaging path during normal system operation. Mirror 1417 reflects the radiation from source 1411 along a path 1419.
The radiation then passes through lens assembly 1407 onto detector 1409. Assuming a uniform detector response (e.g., sensitivity) as is obtainable with a CCD array, and also assuming uniformity in mirror 1417, lens 1415, and source 1411, any non-uniformity measured by detector 1409 is due to lens assembly 1407. The apparatus illustrated in Fig. 14 assumes that lens assembly 1407 exhibits a spherical symmetry, thus allowing the lens non-uniformities to only be determined along a single axis 1421. If lens assembly 1407 does not exhibit such symmetry, additional non-uniformity measurements must be made along other axes.
Another method of characterizing the lens assembly non-uniformities is to remove the lens assembly and characterize it separately from the electrophoresis apparatus. For example in the system illustrated in Fig. 15, a lens assembly 1501 is removed from an electrophoresis apparatus (not shown) and placed between a source 1503 and a detector 1505 on a calibration bench 1507. Given this environment, an extremely uniform source 1503 and detector 1505 may be used thus insuring a high degree of characterization accuracy. Once characterized, lens assembly 1501 is refitted to the electrophoresis apparatus.
Fig. 16 is an illustration of the major components of a system according to the present invention. In this embodiment the key system components are coupled to a
processor 1601 which is coupled to a user interface 1603. Processor 1601 is required for the interpolation and normalization processes to be performed in a timely fashion. This embodiment can be used regardless of the method used to characterize the non- uniformities of lens assembly 1605. To take a sample image, a source 1607 is turned on causing the appropriately marked regions of sample 1609 to fluoresce. The fluorescence is imaged by lens assembly 1605 onto a detector 1611. The sample image may either be displayed immediately on a display monitor 1613 or stored as a sample file within a memory associated with processor 1601. In order to correct the sample image for lens non-uniformities, the lens aperture and magnification settings used during the acquisition of the sample image must be input into processor 1601. These settings are required in order to allow processor 1601 to determine the appropriate correction file to be used during the normalization process. After the image file is normalized using the appropriate correction file, a corrected image file can be displayed on monitor 1613. A corrected image file can also be stored within memory for later retrieval and use.
There are a number of different ways in which the lens settings may be obtained by processor 1601. The most straightforward method is for the user to manually input the setting information using interface 1603. In a second method, the lens settings are automatically determined by processor 1601. In this embodiment lens 1605 contains one or more setting sensors 1615. Sensors 1615 are coupled to processor 1601 by a line 1617. Sensors 1615 are either mechanical or electro-optical, employing well known technologies.
Fig. 17 is an illustration of another technique utilizing a single calibrated light source to determine the aperture and magmfication settings of lens 107. The light source can be a light emitting diode (LED) or other form of easily calibrated source. The source can be located at a position 1701 or at a stationary, out-of-field position, 1703. If the source is placed within the field-of-view of the imager, such as location 1701, it must be removable. For example, a stage 1705 can be used to move the source into position. In order to use this method to determine the lens settings, the calibrated source is placed into position and turned on. An image is taken of the source using the same settings as were used to take an image of the sample in question. The magnification
setting is determined by looking at the image size of the calibrated source while the aperture setting is determined by the intensity of the source.
Fig. 18 is an illustration of another embodiment of the invention. A sample plane 1801 is illuminated with a light source 1803. In this embodiment light source 1803 consists of a series of light bulbs 1805 although the invention is equally applicable to other light sources. In normal usage a sample, for example an electrophoresis gel, would lie within sample plane 1801. The fluorescence from the excited regions of such a sample would be imaged by a lens assembly 1807 onto a detector 1809. Preferably detector 1809 is a two dimensional CCD array. In order to remove the non-uniformities arising from illumination source
1803, a mirror 1811 is used to sample a portion of the source. In one embodiment of the invention, mirror 1811 is coated with a partial reflector, thus allowing a portion of the light from source 1803 to reach lens 1807 and detector 1809. In this embodiment a second portion of the light from source 1803 is reflected along a path 1813. Assuming that only a small portion of the light is reflected along path 1813 by mirror 1811, one of the principal benefits of this embodiment is that mirror 1811 need not be removed during the normal operation of the system. However, in this configuration mirror 1811 will introduce non-uniformities which also must be taken into account during the correction process. In an alternate embodiment, mirror 1811 is coupled to a stage (not shown). During normal operation the stage moves mirror 1811 out of the light path, insuring an unobstructed light path from a sample at plane 1801 to imaging lens assembly 1807. In order to determine the non-uniformities of light source 1803, the stage moves mirror 1811 into a sampling position such as that shown in Fig. 18.
In order to determine the non-uniformities of illumination source 1803, mirror 1811 reflects a portion of source 1803 along path 1813 through a lens 1815. Lens 1815 focuses the light onto a linear detector array 1817. Preferably detector 1817 is a linear CCD array. However, other forms of in-line detector arrays may be used. Furthermore it is not necessary that array 1817 be continuous. Rather, array 1817 may consist of a series of point detectors and the uniformity profile between point detectors interpolated from the measured points.
The dimensions and location of mirror 1811, lens 1815, and detector 1817 are driven by the symmetry of source 1803. For example, a source such as that shown in Fig. 18 can be expected to be relatively uniform along an axis 1819 due to the
illumination uniformity of extended bulbs 1805 along their individual major axes. In contrast, source 1803 can be expected to have an illumination profile similar to that illustrated in Fig. 3 along an axis 1821. Therefore by measuring the profile of source 1803 along axis 1821 as shown, the non-uniformities of the entire source can be determined.
Once the source non-uniformities have been determined the image of the sample may be corrected. Fig. 19 is a block diagram outlining the principal steps in correcting an image using the present invention. First, the profile of the source must be measured using the technique described above (step 1901). Once the profile has been determined, a correction file is stored (step 1903) within a memory associated with a system image processor.
The correction file should contain the same number of correction data points as the number of points resulting from the sample imaging process. Since the number of correction data points are typically substantially less than the number of sample points, the correction data points must be filled-in by the processor. The filled-in correction data points take into account both the symmetry of the source and the number of pixels in detector 1817. Fig. 20 illustrates the interpolation process. A matrix 2001 represents a 11 by 11 matrix of correction data points, thus reproducing a sample image of 121 sample data points. Assuming a source such as that illustrated in Fig. 18 which is uniform along an axis 1819 and non-uniform along an axis 1821, most of the correction data points can be filled-in by the processor. In the illustrated example, correction data points 2003 of column 2005 were measured using the invention. Correction data points 2007 of column 2005 were filled-in through interpolation. Given the uniformity of the source along axis 1819, remaining columns 2009 are simply reproductions of column 2005.
After the source profile has been measured, the sample image can be taken (step 1905) and stored (step 1907). Prior to normalizing the image, in the preferred embodiment the correction file is first smoothed (step 1909) using a smoothing function such as those well known in the art. The smoothing function insures that any source of noise in the correction file is not amplified by the normalization process, resulting in a flawed, corrected sample image. Lastly, the sample file is normalized by dividing it by the correction file (or smoothed correction file if step 1909 is applied) in order to achieve
a corrected sample image file (step 1911). The corrected sample image file can either be displayed (step 1913) or stored (step 1915) for later use.
Fig. 21 is an illustration of the major components of a system according to another embodiment of the invention. In this embodiment the key system components are connected to a processor 2101 which is connected to a user interface 2102. Not only is processor 2101 required for the interpolation/normalization processes to be performed in a timely fashion, but it also permits much of the process to be automated. In use, the sample image may be taken and stored either before or after the illumination source is profiled. Furthermore, it is not necessary to profile the source for every sample run since the source is relatively constant over short periods of time. Profiling the source for every sample run does, however, improve the correction accuracy. In the preferred embodiment the source is only periodically profiled, the new profile being compared to the last profile to determine if there have been any changes, for example changes due to the aging of a light bulb. Assuming that the sample image is taken prior to the source profile, a sample 2103 is first placed within electrophoresis apparatus 2105. In this embodiment source profile mirror 2107 is attached to a retractable stage 2109. During the sample run stage 2109 is retracted so that mirror 2107 is in a location 2111.
To take a sample image, a source 2113 is turned on causing the appropriately marked regions of sample 2103 to fluoresce. The fluorescence is imaged by a lens assembly 2115 onto a detector 2117. The sample image is stored as a sample file within a memory associated with processor 2101. If desired, the sample image may be displayed on a monitor 2119 prior to correction.
To take a source profile, sample 2103 is removed from apparatus 2105 and mirror 2107 is moved into place with stage 2109. Source 2113 is turned on and a source profile is imaged onto a detector array 2121 with a lens 2123. If necessary, depending upon the symmetry of source 2103, additional source profiles may be taken. For example, a second source profile could be taken along an axis perpendicular to the first by replicating mirror 2107, lens 2123, and detector 2121 along the second axis. Once the source profile or profiles are taken, processor 2101 can create a correction file and if desired, a smoothed correction file. This data is then used to normalize the image file as described above.
In an alternate embodiment of the invention the source profiling mirror can be used to profile the lens assembly. This embodiment is illustrated in Fig. 22. This embodiment can be used in conjunction with the previously described embodiment in order to obtain a sample image corrected for both source and lens non-uniformities. If the source is a uniform source this embodiment can be used separately in order to correct for non-uniformities in the lens assembly alone. The source may be made uniform using a variety of techniques such as scanning sources, reflectors, and/or masks.
In order to remove non-uniformities due to lens assembly 1807, a secondary source 2201 is used. Source 2201 is a line source, for example a series of light emitting diodes (LEDs) 2203. Source 2201 may also consist of a series of other types of point sources or a single line source such as a long light bulb. The radiation from source 2201 is enlarged with a lens 2205 before being reflected by mirror 1811. As in the previous embodiment, mirror 1811 may be a partial reflector thus allowing it to remain in place during the normal operation of the system. Mirror 1811 may also be dedicated to the task of measuring lens and/or illumination source non-uniformities. If mirror 1811 is a dedicated mirror, it is placed on a stage so that it can be moved out of the sample imaging path during normal system operation.
Mirror 1811 reflects the radiation from source 2201 along a path 2207. The radiation then passes through lens assembly 1807 onto detector 1809. Assuming a uniform detector response (e.g., sensitivity) as is obtainable with a CCD array, and also assuming uniformity in mirror 1811, lens 2205, and source 2201, any non-uniformity measured by detector 1809 is due to lens assembly 1807.
The apparatus illustrated in Fig. 22 assumes that lens assembly 1807 exhibits a spherical symmetry, thus allowing the lens non-uniformities to only be determined along a single axis 2209. If lens assembly 1807 does not exhibit such symmetry, additional non-uniformity measurements must be made along other axes.
Once the lens non-uniformities have been determined, the sample image may be corrected following the same steps as outlined in Fig. 19. The only difference is that instead of measuring the profile of the source, step 1901 is measuring the profile of lens assembly 1807. The remaining steps such as creating a correction file and then using the correction file to normalize the sample file are the same as previously described.
Fig. 23 is an illustration of a system according to the present invention which includes the components necessary to perform the lens non-uniformity
measurements. This figure is identical to that shown in Fig. 21 with the addition of source 2201 and lens 2205. In this embodiment source 2201 is connected to processor 2101, thus allowing the lens non-uniformity measurement to be automated.
Fig. 24 is a block diagram outlining the principal steps used in accordance with another embodiment of the invention. In this embodiment the profile of the source is first determined using the system described above (step 2401). Once the source profile is determined, the profile of the secondary source can be altered to mimic the measured profile of the actual source (step 2403). This step requires that source 2201 be comprised of a series of individually controllable point sources 2203. Preferably point sources 2203 are pre-calibrated and controlled by processor 2101. Once the profile of secondary source 2203 has been properly adjusted, a lens profile can be measured (step 2405) and a correction file stored (step 2407). Since source 2201 has been regulated in order to provide the same profile as source 2403, the stored correction file takes into account both the non-uniformities in illumination source 2403 and the non-uniformities in lens assembly 2407. As in the previous embodiments, the correction file must contain the same number of correction data points as the number of data points resulting from the sample imaging process. Thus the correction file contains a number of correction data points which are not measured, rather they are computed by processor 2101 using a standard interpolation function. After the correction profiles have been measured and an appropriate correction data file has been stored, the sample image is taken (step 2409), stored (step 2411), and normalized (step 2413). As in the previous embodiments, if necessary a smoothing function can be applied to the correction data file before it is used to normalize the sample image. Lastly the corrected sample file is either stored (step 2415) for later use, or immediately displayed (step 2417).
As will be understood by those familiar with the art, the present invention may be embodied in other specific forms without departing from the spirit or essential characteristics thereof. For example, this invention is not limited to specific excitation and emission wavelengths. As such, UV excited markers which fluoresce in the visible wavelengths may be used. Also, other techniques of characterizing the lens non- uniformities may be used. Accordingly, the disclosures and descriptions herein are intended to be illustrative, but not limiting, of the scope of the invention which is set forth in the following claims.