CN101002093A - Image analysis and assay system - Google Patents

Abstract

Systems for determining and/or analyzing the distribution and dynamics of cellular components.

Description

Graphical analysis and pilot system

The cross reference of priority application

The application is according to the rights and interests that also require according to the following U.S. temporary patent application of 35 U.S.C. § 119 (e), and its integral body with them is incorporated herein by reference and is used for all purposes: sequence number No.60/537, and 454, submit on January 15th, 2004; With sequence number No.11/039,077, to submit on January 17th, 2005, name is called IMAGE ANALYSIS SYSTME, and names Vladimir Temov and Ilya Ravkin as the inventor.

The cross reference of related application

The application is incorporated herein by reference the integral body of following U.S. Patent application to be used for all purposes: sequence number No.09/549, and on April 14th, 970,2000 submitted to; Sequence number No.09/694, on October 19th, 077,2000 submitted to; Sequence number No.10/120, on April 10th, 900,2002 submitted to; Sequence number No.10/238, on September 9th, 914,2002 submitted to; Sequence number No.10/273, on October 18th, 605,2002 submitted to; Sequence number No.10/282, on October 28th, 904,2002 submitted to; Sequence number No.10/282, on October 28th, 940,2002 submitted to; Sequence number No.10/382, on March 5th, 796,2003 submitted to; Sequence number No.10/382, on March 5th, 797,2003 submitted to; Sequence number No.10/382, on March 5th, 818,2003 submitted to; Sequence number No.10/407, on April 4th, 630,2003 submitted to; Sequence number No.10/444, on May 23rd, 573,2003 submitted to; Sequence number No.10/445, on May 23rd, 291,2003 submitted to; Sequence number No.10/713, on November 14th, 866,2003 submitted to; Sequence number No.10/842, on May 10th, 954,2004 submitted to; Sequence number No.10/901, on July 28th, 942,2004 submitted to; With sequence number No.10/942, on September 15th, 322,2004 submitted to.

The application is incorporated herein by reference the integral body of following U.S. Provisional Patent Application to be used for all purposes: sequence number No.60/129, and 664,1999 year April 15, order was submitted to; Sequence number No.60/170, on Dec 15th, 947,1999 submitted to; Sequence number No.60/241, on October 18th, 714,2000 submitted to; Sequence number No.60/259, on Dec 28th, 416,2000 submitted to; Sequence number No.60/293,863, submit to May 24 calendar year 2001; Sequence number No.60/299,267, submit to June 18 calendar year 2001; Sequence number No.60/299,810, submit to June 20 calendar year 2001; Sequence number No.60/307,649, submit to July 24 calendar year 2001; Sequence number No.60/307,650, submit to July 24 calendar year 2001; Sequence number No.60/310,540, submit to August 6 calendar year 2001; Sequence number No.60/317,409, submit to September 4 calendar year 2001; Sequence number No.60/318,156, submit to September 7 calendar year 2001; Sequence number No.60/328,614, submit to October 10 calendar year 2001; Sequence number No.60/343,682, submit to October 26 calendar year 2001; Sequence number No.60/343,685, submit to October 26 calendar year 2001; Sequence number No.60/344,482, submit to October 26 calendar year 2001; Sequence number No.60/344,483, submit to October 26 calendar year 2001; Sequence number No.60/345,606, submit to October 26 calendar year 2001; Sequence number No.60/348,025, submit to October 26 calendar year 2001; Sequence number No.60/348,027, submit to October 26 calendar year 2001; Sequence number No.60/359, on February 21st, 207,2002 submitted to; Sequence number No.60/362, on March 5th, 001,2002 submitted to; Sequence number No.60/362, on March 5th, 055,2002 submitted to; Sequence number No.60/362, on March 5th, 238,2002 submitted to; Sequence number No.60/370, on April 4th, 313,2002 submitted to; Sequence number No.60/383, on May 23rd, 091,2002 submitted to; Sequence number No.60/383, on May 23rd, 092,2002 submitted to; Sequence number No.60/413, on September 24th, 407,2002 submitted to; Sequence number No.60/413, on September 24th, 675,2002 submitted to; Sequence number No.60/421, on October 25th, 280,2002 submitted to; Sequence number No.60/426, on November 14th, 633,2002 submitted to; Sequence number No.60/469,508,2003 year May 8, order was submitted to; Sequence number No.60/473, on May 22nd, 064,2003 submitted to; Sequence number No.60/503, on September 15th, 406,2003 submitted to; Sequence number No.60/523, on November 19th, 747,2003 submitted to; With sequence number No.60/585, on July 2nd, 150,2004 submitted to.

The application is incorporated herein by reference the integral body of following PCT patented claim to be used for all purposes: sequence number No.PCT/US01/51413, submit to October 18 calendar year 2001, and open to disclose No.WO 02/37944 on May 16th, 2002.

Introduce

The machineization of molecule and supermolecule assembly and dynamics play important effect in the function of cell system.Especially relevant on eukaryotic and the many structures and/or on the function, ad-hoc location or such as the organising of composition height in the organelle official rank compartment organises.For example, selection with eukaryotic in energy produce relevant cell component and organise in the mitochondria, and organising in the nucleus of selecting with cell control and hereditary relevant cell component.More common, eukaryotic can comprise many different machineizations being used for many different functioning cell organ or compartments, and it comprises nucleus, mitochondria, chloroplast, lysosome, peroxisome, vacuole, golgiosome, coarse or smooth endoplasmic reticulum, centrosome, plasma membrane, nuclear envelope, endosome, excretion vesicles or the like.

The ingredient of these different compartments and cell and biologic artifact can be high power attitude usually.Therefore, specific molecular can the zones of different in cell between and/or active transhipment and/or diffusion between cell and extracellular medium.In some cases, cell can move or transposition to another compartment from a compartment, with variation, cellular signal transduction (for example hormone), morbid state of response cell cycle or the like.In addition, under equimolecular situation such as enzyme, control the mechanism that these distribute and dynamic (dynamical) mechanism can be independent of control or influence catalytic action, this means that they can provide target unique, that do not utilize before to drug candidate, it is potential allows to have the functional composition of identity function (such as kinases) according to different position or transposition signal or behavior and by target.Noticeable, a lot of potential molecules relevant with morbid state, such as transcription factor and kinases, in the activation step process especially from tenuigenin to the nucleus transposition.

Be to be a plurality of compartments such as " nature " method of transposition graphical analysis etc. in the graphical analysis with image segmentation, nucleus and tenuigenin such as separate cell, the quantity of the signal dyeing of measurement in each compartment is also calculated the tolerance (1,2) as the transposition of ratio of the two or difference.The variant of this method is the signal dyeing of analyzing in the smaller area chamber, and described compartment defines (3,4) by the spatial relationship at they and nuclear border or center.In all cases, these methods need image segmentation.Therefore, because cut apart usually to characteristics of image and pseudomorphism (artifact) sensitivity, and may be not goodly with the magnification convergent-divergent, this just needs such system, and described system does not need or undemanding at least depending on cut apart.

General introduction

This theory is provided for measuring and/or analysis of cells component distributing and dynamic (dynamical) system.

The accompanying drawing summary

Fig. 1 is the synoptic diagram according to the general framework that is used for graphical analysis of this theory.Picture A represents to have the cell of different reporter images, and picture B represents exemplary 3-D histogram, and picture C represents three kinds of exemplary two-dimensional histograms,

Fig. 2 is a series of microphotos, and it is illustrated in the nuclear transposition of NF κ B in the MCF7 cell.The state of a left side-feminine gender, centre-centre, the right side-positive.The image of FITC dyeing obtains with the 10X object lens.

Fig. 3 is an exemplary dyeing of cover and the counterstain distribution plan by model and true cell.A, B-model; C, D-are real; A, the C-positive; B, the D-positive.The S-simple stain; The CS-counterstain.

Fig. 4 is a simple stain (Z-axis) in desirable model system (A, B), interference model system (C, D) and true cell (E, F) and a cover intersection histogram of counterstain (transverse axis).Scale on two axles is 0-255.Line is represented the approximate value calculated.

Fig. 5 is a cover chart, and the approximate value of its expression slope (left side) is as the function of the counterstain intensity in the subclass (right side) of the distribution that increases from right to left.Center line is represented the slope value of gained.Top picture-protein appraise and decide the position; The tenuigenin location of end picture-protein.

Fig. 6 is used for the possible numerical procedure process flow diagram of tenuigenin to the graphical analysis of nuclear easy bit test.

Fig. 7 is independent (cell (cell-by-cell) one by one) and the chart of global slopes of the different magnifications of a cover expression.

Fig. 8 represents that how changing one of two-dimentional dye distribution by the generation point group by the kernel shown in the ellipse of dotted line overlaps chart, and it causes underestimating slope.This influence can be proofreaied and correct corresponding to the hole in the signal pattern of kernel by filling.A, B-single cell analysis; C, the holistic approach of D-many cells.A, C-be correction data not; B, D-proofread and correct (filling) data.Slope: (A) 61, (B) 72, (C) 1.21, (D) 1.61.

Fig. 9 is the exemplary method a series of picture that be used to fill kernel of expression according to this theory each side.

Figure 10 is that the tenuigenin of the NFKB in the MCF7 cell overlaps a cover histogram of the slope distribution in the relevant image of dosage to one of the transposition of examining.This histogram represents to have the number percent of given slope (Z-axis) to the cell of slope (transverse axis).

Figure 11 is the bar graph of expression by the variation of the data set of histogrammic first principal component explanation of slope.

Figure 12 is the chart of the principal component weight on the primitive character unit (bins) that is illustrated in the slope histogram.

Figure 13 is that tenuigenin in the space of expression slope histogrammic first liang of principal component is to the curve map of the image distribution of nuclear easy bit test.Dotted arrow is represented the increase of TNF α dosage.

Figure 14 is the dose curve of nuclear transposition, and expression average cell slope (Z-axis) is to TNF α concentration (transverse axis).The V-factor=0.77.

Figure 15 is the chart of the V-factor, and it is used for the tolerance (Z-axis) of conduct in the nucleus transposition of the function of the interpolation magnification (transverse axis) of different images size.

Figure 16 is a series of microphotos, and its expression film is to cytoplasmic transposition.The state of a left side-feminine gender, centre-centre, the right side-positive.

Figure 17 is for being used for the cover chart of film to the nucleus counterstain and the monochromatic joint distribution of the model system of tenuigenin transposition.Top-film location, bottom-tenuigenin location.The distribution plan of the Two dimensional Distribution of a left side-dyeing, the right side-by model cell.

Figure 18 is a cover chart that is used in the joint distribution of true cell simple stain and counterstain.Top-film location, bottom-tenuigenin location.The distribution plan of the Two dimensional Distribution of a left side-dyeing, the right side-by model cell.

Figure 19 is for being used for the cover chart of film to the joint distribution of the model system simple stain of disturbing with random noise of tenuigenin transposition and film counterstain.Top-film location, bottom-tenuigenin location.The distribution plan of the Two dimensional Distribution of a left side-dyeing, the right side-by model cell.

Figure 20 is to be a series of images of the Transfluor  test that has 2*2 dividing elements (binning) of 10X at objective lens magnification.The A-feminine gender, that B-is middle, the C-positive.

Figure 21 is used for cover intensity map Figure 24 test, by cell in original image, in by the image of size 1 structural element opening operation and in the image by the structural element opening operation of size 4.The state of a left side-feminine gender, centre-centre, the right side-positive.

Figure 22 is a cover curve map of the feminine gender of expression Transfluor test, granular spectrum middle and positive.Transverse axis-opening operation size, Z-axis-at the percentage of the image volume of this opening operation.

Figure 23 is that expression is used at the dependent cover curve of the magnification (transverse axis) and the z value (Z-axis) of the relative granularity of image size (as shown in the figure).The preferably profile of the scope of test performance has drawn.

Describe in detail

This theory provides distribution and the dynamic (dynamical) system that is used for mensuration and/or analysis of cells composition. These can comprise for the preparation of the device of, location, treatment and/or analyzing samples etc., method, The system of composition and kit can be particularly useful for grinding of two or more material Joint Distribution Study carefully, especially wherein these materials of one or more play the effect of reference or counterstain, and one Kind or more these materials play the effect of signal dyeing. For example, in some embodiments, reference Or counterstain can be with the mark that acts on cell characteristic or compartment, and signal dyeing can be used for energy The research of the distribution of enough materials with the cell transposition. These transpositions can comprise that cytoplasm is to cell The transposition, nucleus of nuclear to cytoplasmic transposition, film to the transposition of cytoplasm (or nucleus), carefully Kytoplasm (or nucleus) arrives transposition of film etc.

As using preparation sample herein to comprise, wherein, (1) selects, separates, concentrates, cultivates, modifies and/or synthetic composition, cell component, cell, tissue and/or any other are tested to become to grade, (2) select, form and/or modify sample carriers and/or sample container, respectively such as code carrier and/or porous system, with such as microtest plate, and/or (3) sample and sample carriers and/or sample and sample container etc. of being associated.

Can comprise as location as used herein sample and to be used for the treatment of and/or the location sample (and/or any relevant sample carriers) of analysis etc.Such location wherein can comprise, (1) mixing sample, and (2) distribute sample in treatment and/or site of analysis, and/or (3) disperse sample in treatment and/or site of analysis, for example, allow visual near sample and/or sample respectively.

Can comprise as treatment sample as used herein sample is exposed to some conditions, and/or similarly and/or its variation such as chemicals, temperature, the concentration ion concentration of hydrogen ion (pH), salt ion etc. (for example, such as).These conditions can comprise candidate's adjusting, and condition that is for example unknown or the Partial Feature effect is such as candidate's transcriptional regulatory.

As analyzing samples as used herein can comprise qualitatively and/or quantitative observation and/or measure sample condition (for example size, quality, density etc.) and/or the condition that causes by sample (for example, the loss of enzyme matrix, the generation of enzyme product etc.), use any suitable method (for example, (imaging, absorption, scattering, luminous, photoluminescence (for example fluorescence or phosphorescence), chemiluminescence etc.), magnetic resonance and/or fluid mechanics of optics etc.).Such analysis can also comprise that detection and/or explanation comprise existence, quantity and/or the activity of sample or its adjusting of excitant and/or antagonist, and/or measures trend or purport from multiple sample analysis.Such analysis can also comprise measure and/or analysis of biological system in the joint distribution of two or more dyeing or other position and/or activated indicators, for example be used for easy bit test etc.

The system that provides by this theory also comprises, but be not limited in following examples, describe, and optionally coupling apparatus, method (comprising mark and transfection method), composition (comprising molecule, cell, tissue etc.) and/or kit or its component, described in different patented claims listed in above cross reference, and be incorporated herein by reference.

Embodiment

Following embodiment has described the aspect and the embodiment of the selection of this theory, especially exemplary distribution and dynamic test.These embodiment are comprised being used to illustrate the entire scope that is not intended to limit or define this theory.The other aspect of this theory is described in the different patented claims of above listed cross reference and is incorporated herein by reference, especially the U.S. Provisional Patent Application sequence number N.60/537,454, submit on January 15th, 2004; With U.S. Provisional Patent Application sequence number No.____, to submit to January 17 in 2005, exercise question is IMAGE ANALYSIS SYSTEM, and names Vladimir Temov and Ilya Ravkin as the inventor.These two temporary patent applications comprise coloured picture and text extra, that replenish and further specify the following stated notion, especially in embodiment 1,2 and 4.

Embodiment 1. Tenuigenin is to the test of nuclear transposition

1.1. background

Fig. 1 illustrates the general data framework of this theory exemplary embodiment: tenuigenin is to the test of the transposition of nucleus (or nucleus is to tenuigenin).Visual field (field of view) is in different spectral ranges and/or with (or obtaining that different optical modes digitally obtains with simulating, and convert to the numeral), so exist, maybe can make it becomes from the unanimity of the pixel in all images of homologous field to pixel.Different spectral ranges can comprise different wavelength bands, such as blue and green etc.The different optical pattern can comprise different imaging techniques, such as photoluminescence and transmission etc.This framework allows the analysis of Two dimensional Distribution or a series of this distribution or up to the analysis of the more high-dimensional distribution of the number of report image.Such joint distribution can be analyzed in the different subclass of pixel, scope from entire image down to single celled part.

1.2 Method based on the 2D distribution of dyeing.

Model and experiment distribute.

This theory can comprise the analysis according to the transposition incident of the joint distribution of signal and counterstain.Collect representative data and at MCF7 transit cell record factor NFkB in response to the transposition of TNF α concentration (seeing for example Fig. 2) to its analysis.In order to find the stable tolerance of nucleus transposition, we have also defined and have studied the model of the space distribution of nucleus counterstain and signal dyeing, when it from tenuigenin to the nucleus transposition time.The model result of comparing with experimental data is studied to seek stable tolerance in different condition with under disturbing.

Fig. 3 represents the intensity distribution along line, and described line passes model (picture A and B) and true (picture C and the D) cell that comprises signal dyeing and counterstain.Model cell comprises bell type counterstain (nucleus) intensity distributions, with or (picture A) bell type signal staining power of having a broad of bell type depression distribute, corresponding to the negative correlation between signal and counterstain, or the distribution of (picture B) bell type signal staining power, corresponding to the positive correlation between signal and the counterstain.True cell presents the distribution plan similar substantially to model cell.Herein, drafting presents negative correlation by the distribution plan of the picture C of two cells, and drafting presents positive correlation by the distribution plan of the screen D of three cells.All models and real distribution plan are arrived their maximum of intensity separately by independent standardization.

1.3 The quantification of crosscorrelation

Use any suitable method can observe and/or analytic signal dyeing and counterstain between and/or the crosscorrelation or the joint distribution of its variation.In some cases, visually observed value or variation are possible and enough simple.But in most of the cases, observed value or variation are desired or necessary quantitatively, especially in the background such as the scanning that may comprise many sample analyses.

The intersection histogram or the correlation graph of Fig. 4 shows signal dyeing (Z-axis) and counterstain (transverse axis).Concrete, the intensity (or its some suitable amount degree or functions) of signal dyeing is plotted as the function of relevant counterstain (or its some suitable amount degree or functions) intensity.Therefore, data point in the left lower quadrant of chart is corresponding to having the two the image section of low concentration of signal dyeing and counterstain, data point in the right upper quadrant is corresponding to having the two high-density images zone of signal dyeing and counterstain, data point in the left upper quadrant is corresponding to having high signal dyeing concentration, but the image-region of the concentration of low counterstain, data point in the right lower quadrant is corresponding to having the low signal dyeing concentration, but the image-region of the concentration of high counterstain.In these charts, negative correlation will trend towards being disclosed as the distribution of data points with negative slope, and positive correlation will trend towards being disclosed as the distribution of data points with positive slope.In order to obtain stable tolerance, described tolerance is being converted to feature from negative to positive situation, or vice versa, and we analyze the joint distribution of the dyeing on model and true cell.In ideal conditions, model space dye distribution circulation symmetry and calibration, illustrate as Fig. 3 (A, B).Intersection histogram such as Fig. 4 (A, B) of being used for these situations are shown.If the center by two dyeing of setovering, by changing shape and/or increase noise etc. and interference model from the circle to the ellipse, it is fuzzy that this distribution becomes, shown as Fig. 4 (C, D).Common feminine gender and positive true cell have the shown intersection histogram as Fig. 4 (E, F).These distribute and show that indirectly suitable transposition tolerance can be defined as the slope (or symbol of more coarse described slope) of the straight line line segment that is similar to the histogram right side.This part distributes corresponding to stronger nucleus dyeing and also near nuclear center.Decentering is far away more, and described distribution is loose with regard to overstepping the bounds of propriety, and the reliability of approximate value is few more.

Fig. 5 shows the function of the approximate value of slope (left side) as the counterstain intensity in the subclass (right side) of the distribution that increases from right to left.By drawing approximate slope from right to left, and select the scope at stable place of approximate value, and obtain being used for the distribution part approximate with straight line.

Fig. 5 also is presented at may change on this method.This variation can comprise calculates two slopes again.Top line is the tropic that calculates with all points on initial slope segment (we will be referred to as slope 1); Bottom line is the tropic that calculates with all points below initial slope segment (we will be referred to as slope 2).If all three slopes (that is, initial slope, slope 1 and slope 2) have identical symbol, the result has of maximum value so.But,, select original slope so if they have different symbols.We claim this tolerance to be slope 3.

1.4 Whole, one by one cell, group's analysis one by one

This theory can be applied to entire image or its part, includes, but are not limited to the part of single celled selection, the group of the cell of the cell of selection and/or selection or zone etc.Fig. 6 demonstration is used for the process flow diagram of tenuigenin to the possible numerical procedure of the graphical analysis of the easy bit test of nucleus.

The variation of the method in the application of unicellular level can offset or be compressed on expression or dye, it can be non-information under the situation of transposition.Certain situation (for example, low magnification), with image segmentation be unicellular be difficult; Can on the cell mass of close positions, carry out and analyze.This may not illustrate the biomutation between the cell in the group, but it will illustrate the variation between the group.Variation between the group can also since technology or the experiment, such as uneven illumination.Analysis can be applied to unicellular, does not need the understanding on cell or nucleus border, but simply need comprise the understanding in the zone of isolated cell.

Holistic approach also has its advantage.It can be faster and/or more stable at low magnification.The objection of analyzing at whole (whole) is that usually it can not illustrate intercellular variation, also is not get rid of unwanted cells.In any case analyze the cell accepted, independent or as a whole, second problem can directly be handled.The purpose that is used for this discussion has 2 points: (1) holistic approach can not provide the tolerance of the same good average response with single cell analysis, and (2) average tolerance may not be full information alone.First can be to the overcoming of small part, and by with strength criterionization, whole in this case tolerance is the same good with the average of unicellular tolerance usually, sees Fig. 7.Second is discussed in 1.9 parts.

1.5 Be divided into component. mark

The watershed divide of bond strength image

Image can be used as integral body and/or part or becomes to assign to analyze.Be divided into part and can be used for two purposes: (1) is convenient to the analysis of selected characteristics of image, such as cell mass, unicellular and/or its part; And (2) are convenient to optionally intensity equilibrium as a step in the program.

Cut apart and can use any suitable mechanism to carry out, such as: (1) seeks mark and (2) seek cut-off rule.

Mark can be sought by any suitable algorithm.For example, fixed value (mark contrast) can be saturated from nucleus counterstain figure image subtraction, and the gained image is rebuild [11] within nucleus counterstain image.And this image can be from counterstain figure image subtraction, and be converted to bianry image.The ingredient of this bianry image is a mark.Can apply further restriction on the mark: the mark that only has at least one pixel on given threshold value (index intensity) is preserved for second step.According to magnification and noise level, the image of the multiple color of nucleus can be smoothed before this algorithm.The method of this mensuration mark can be handled the cell of different sizes and shape.Other method for example also can be used based on the method for cap conversion (top-hat transform) [11].

Cut-off rule between ingredient (for example, nucleus, cell etc.) can be found out by any suitable algorithm.For example, cut-off rule can be defined as the watershed divide [5,6,10] of inverted image image of the linear combination of counterstain image and signal colored graph picture.Use linear combination and just the reason of nucleus counterstain image be frequent asymmetric and uneven separating of cell.Cut-off rule from the nuclear staining image may be by cutting in the middle of the cell.The application of signal dyeing produces more accurate cut-off rule.The coefficient of linear combination can be according to dyeing and the characteristic of Image Acquisition and different.

1.6 Strength criterionization

The joint distribution of counterstain and signal dyeing can be standardized as their maximal values separately.This can or finish on initial pictures on distributing.The result is identical, but makes image standardization provide extra feedback to the user, and may be displayed on standardization sightless feature before.

Use any suitable mechanism, standardization (and/or other rescaling) can be carried out on entire image and/or its part.For example, as mentioned above, standardization can be carried out in component.In this situation, all pixels from a component multiply by identical numeral, are respectively applied for signal dyeing and are used for counterstain.Interchangeable, by smooth surface being applied to the image of signal dyeing and counterstain, standardization can be finished and not need split image.Standardization can have the effect of partial equilibrium image, and can comprise that graphical rule changes, and so maximal value and/or integration value equal unit value or some other preset values.

1.7 The removal of pseudomorphism, gate, classification

Can the exert an influence pseudomorphism of test findings of physiological changeability and/or other condition.Some cells for example, such as the MCF7 of enough low density, it has the number percent of conspicuous mitotic cell, and its nuclear membrane is broken and chromosome is assembled.Chromosome can produce " feminine gender " result of forgery with these cells of the intensive dyeing of nucleic acid dye and upset the positive of test.But these cells can be excluded (or removal) according to their high nuclei dyeing intensity of colour and/or " nucleus " etc. that is significantly less than normal size (undersized).Herein, " eliminating " can comprise calculating and/or the tabulation that is not used in subsequently, and/or is not used in mensuration of last test findings etc.Described information that can be excluded or result can comprise the part and/or the integral body of one or more cells, one or more cell compartment etc.Therefore, in exemplary embodiment, wherein cell and fluorescence wire connection are touched, or extend down thereon or at it, and the part that cell is affected can be excluded, and/or all cells that are affected can be excluded etc.More generally, also can be excluded such as any pseudomorphism other cell type and/or acellular pseudomorphism etc., that can be distinguished by its intensity, shape, size and/or position etc.

Opposite, in some cases, such as can being used for cytological classification intersecting the crossing dependency of slope value in the histogram etc., rather than get rid of cell.For example, mitotic cell can raise and intersect negative slope in the histogram, because signal dyeing will trend towards being excluded from multiple (nucleus) pigmented section, and intermitotic cell positive slope that can raise, if there is positive correlation between the position of signal dyeing and counterstain at least.

1.8 Flame Image Process. the kernel by filler opening is removed

Protein from tenuigenin to the nucleus transposition and other molecule do not enter kernel usually.This tendency can produce pseudomorphism, unless consider, because it can be interpreted as the disappearance of transposition.

Fig. 8 is illustrated in about the error in the intersection histogram slope estimation of kernel.Cause that these errors are relevant with the zone of low signal staining power because have the zone of high counterstain (be nucleus dyeing) intensity, or even when having transposition, because dye from kernel eliminating signal.

These pseudomorphisms can by identification kernel and from the nucleus mask, remove they mask and processed.But this method suffers usually and cuts apart the shortcoming identical with mask (scape of passing away).

As if pseudomorphism can also be handled by the image that changes signal dyeing, and so it does not have the feature of not expecting, for example by filler opening, and do not have kernel.The whole nucleus of filling kernel but not filling negative cells is a challenge, and the whole nucleus of described negative cells seems also as the hole.A kind of method is the image of taking advantage of operation based on pixel that (1) generates signal and counterstain image, the hole [10] in (2) blank map picture, and (3) add increment to original signal colored graph picture.This increment can be taken advantage of by the constant greater than 1.The shortcoming of this method is may not have complete filling near the hole (kernel) at nucleus edge.Interchangeable method is a direct filler opening on signal colored graph picture, but only be chosen among them those and fall into (that is, its type, condition setting or the like neither too little neither be too big) as the scope of the size of kernel feature for given cell.

Fig. 9 represents to be used to fill the exemplary steps of kernel.Dotted line in each picture is illustrated in the signal dye distribution figure that imagines under the situation that does not have kernel.Picture A represents the intensity distributions by kernel.Picture B is illustrated in the intensity distributions after the filler opening.Picture C is illustrated in level and smooth intensity distributions afterwards.At last, screen D is illustrated in the image level and smooth and that fill selected under the mask of fill area at utmost.Optionally level and smooth step can also be improved the intensity distributions after the filling.

More common, image can be modified, if this estimation that causes better interested final test to be measured, for example, as the described measured quality of 1.10 parts.An embodiment is level and smooth.In some cases, this can improve slope tolerance, if especially image is acquired having on the equipment of the shortsighted wild degree of depth.

1.9 Xenogenesis. group's tolerance of position and variation

Principal component analysis (PCA)

This theory comprises the system that is used for illustrating or explaining the cell mass xenogenesis.For example, protein from tenuigenin to nuclear transposition process, the action that not all cell is all synchronous, and different cell even can be rendered as opposite behavior.

In certain situation, single (scalar) tolerance of seeking or measuring transposition can be possible or expectation.Under these circumstances, can reasonably the group be reduced to the position and measure, such as average (on average), intermediate value, mode etc.The tolerance that makes a variation in cell mass can also provide valuable information.In appearing at embodiments herein, such as standard deviation, be rendered as the relevant behavior of dosage, exactly be similar to location measurement from the variation tolerance of the median deviation of intermediate value etc.

In identical and/or other situation, multidimensional (vector) tolerance that finds or measure transposition can be possible or expectation.In such example, it can reasonably use the multidimensional statistics method, such as principal component analysis (PCA) [12] (PCA).Multidimensional analysis can provide extra or more detailed information about cell behavior and xenogenesis.

Figure 10 is a cover histogram or a curve map, and it is illustrated in the slope distribution in the cover image, the NF κ B transposition relevant from tenuigenin to nuclear dosage in the described graphical representation MCF7 cell.Herein, the cell number percent (Z-axis) of given slope is expressed as the function (transverse axis) of slope.

The result of data is analyzed in Figure 11-13 expression from the multidimensional PCA among Figure 10.Multi-C vector in the present embodiment is the histogram of slope distribution in the cell, as shown in figure 11.Feature is the unit (bin) in histogram, and data set (cases) is a dosage and in the possible repetition of each dosage.Major component is one group of irrelevant vector, and it is the linear combination of original vector group.According to the character that data are provided with, initial some major components can be explained major part variation wherein.For example, initial two main compositions are herein explained almost 90% variations, are rational so the dimension of data set is reduced to 2 from ten.By analyzing their weights on primitive character, its meaning is given to main composition, and is as shown in Figure 12.In this embodiment, first principal component can be interpreted as the positive of transposition, because the weight of negative slope is born, the weight of positive slope is positive.The second main composition can be interpreted as homogeney, because high just the two all has positive weight with high negative slope, and near the slope zero has negative weight.NF for the MCF7 cell _κThe dose curve of the nuclear transposition of B and the distribution of point may be plotted in the space of histogrammic two major components at first of slope, as shown in figure 13.

1.10 Be used for the test of cell imaging test and the tolerance of algorithm quality

In the cell imaging test, the tolerance (or a plurality of tolerance) that is used for describing assay features can remove from the signal by camera record.In addition, the different algorithm different test tolerance that on identical image, can produce.(for example, the nuclear transposition) test is especially sharp, and wherein total intensity can be constant, and wherein test findings can rely on algorithm more than depending on original image for redistribution for this.In order to determine which kind of solution is the bottom line acceptable for given test and algorithm, we analyze the identical bore region of different optical magnification or/and the identical image group of different interpolation magnifications.In a similar fashion, by comparing the influence of measuring analysis of cells quantity from different sized images.For comparative result, the quality metric that we use this place to discuss.

Such as the quality test of high production shaker test etc., can calculate by the dynamic range and the variable statistics parameter that depend on test, such as the z-factor [9]:

$Z=1-3\left\{\frac{{\mathrm{SD}}_{\mathrm{pos}}+{\mathrm{SD}}_{\mathrm{neg}}}{|M_{\mathrm{pos}}-M_{\mathrm{neg}}|}\right\}$

Herein, SD is a standard deviation, and M is a mean value, and pos and neg be two kinds of extremities of test, and it defines its dynamic range.The scope of Z factor is from-∞ to 1.Be used for the test based on cell, it is good that the z factor greater than 0.5 is considered to.Z factor has been proved to be to for catching and and being highly effective by test organism and the variability that caused by use instrument (for example, suction amount) relatively.Introduce several new variablees based on the test cell line of imaging: imaging resolution, imaging region size and data extract algorithm.The imaging region size is a variable and since its usually less than the total system hole of whole microtest plate (for example, less than) by imaging and analysis.Owing to have the quality metric that is similar to the z factor, allow us to make variable optimization under our control, for example, find best data extract algorithm.Herein, we will handle specific cytological image analyses algorithm and service property (quality) tolerance with optimization image resolution ratio and size.

Except introducing new variable, the cell imaging test can guide us to rethink quality metric itself.The test that is derived from image measure may be calculate very complicated.For example, they can to comprise that operation has saturated from the feminine gender of test and the effect of the value of positive, artificial minimizing variability.This may take place and even do not recognized unintentionally.In addition, if the value of the test of positive negative status in for it not overlapping (and if they are overlapping, it may not be very useful test), by using all in being mapped to independent numerical value, and the mathematics conversion in the other independent value of all negative values mapping, the z factor can be by autotelic operation.A kind of mode of handling this situation is to use the sequence-dependent quality metric of dosage of trystate (dose curve), has enough approaching each other dosage, and so manually-operated will be impossible.This causes following measuring, and we are referred to as " the v-factor ":

$V=1-6\left\{\frac{{\mathrm{SD}}_{\mathrm{of}\_\mathrm{fit}}}{|M_{\mathrm{pos}}-M_{\mathrm{neg}}|}\right\}$

Herein

${\mathrm{SD}}_{\mathrm{of}\_\mathrm{fit}}=\sqrt{\underset{1}{\overset{n}{\mathrm{\Σ}}}\frac{{(f_{\exp }-f_{\mathrm{mod}})}^2}{n}}$

Herein, f _ExpAnd f _ModBe empirical value and the model numerical value of measuring in the test of given concentration respectively, and n is the number of the experimental point in dose curve.

If two dose points are only arranged, the V factor is converted to the z factor.Can be according to the character preference pattern of reaction, curve often is the selection of nature.Interchangeable, in some cases, do not use particular model, and the mean value of several repetitions is as the f in aforesaid equation _ModYet, provide the v factor by this formula:

$V=1-6\left\{\frac{\mathrm{Average}\_{\mathrm{SD}}_{\mathrm{neg}}}{|M_{\mathrm{pos}}-M_{\mathrm{neg}}|}\right\}$

The influence of the saturated pseudomorphism that the described v factor is caused by calculating than the z factor being subjected to easily still less.The nuance that also has other.Standard deviation in the middle of the dose-effect curve is often greater than in extreme standard deviation.This is that therefore any dispensing error is very little to the influence of reaction because the peak on the curve often is determined as in saturation concentration.Minimum point is zero-dose normally, and it has also avoided dispensing error.Opposite, the influence of volumetric errors has its maximum efficiency in the centre of dose-effect curve.Therefore, at least owing to these reasons, consider that entire curve can provide the measuring of test figure quality of reality more.

1.1 dosage relies on. image size and magnification rely on

The mean value of unicellular slope can be used as test parameters, for example makes the data characterization from the hole.

Figure 14 represents to examine the transposition dose curve, is used to calculate the adaptability as the average cell slope of test parameters.From one group of dose-dependent image collection data, such as those in Fig. 2.Herein, average cell slope (Z-axis) is plotted as the function of TNF α concentration (transverse axis).The corresponding v factor is 0.77.

Figure 15 represent as the function of the interpolation magnification (transverse axis) of different images size (be reported as square millimeter), be used to examine the v factor (Z-axis) that transposition is measured.Concrete, the behavior of this algorithm is studied as (1) interpolation image magnification, drops to the magnification of 2X from 10X magnification originally, and (2) image size, from 0.510mm ²Down to 0.009mm ², function.The v factor is as quality metric.Finish image interpolation by Bilinear Method.In order to study the dependence of image size, be divided into the image segments of reduced size for the original image of each point in the curve.Next, each less image is used to produce transposition to be measured, and these measure the formula that is used for the v-factor.It is shown in Figure 15 to be used for these different results that analyze, the maintenance level state of the v factor of its expression algorithm arrival about 0.8, and it is at 4X or bigger magnification, and the image size is 0.34mm ²Or under the bigger situation.

Average cell slope algorithm can have the feature of several expectations: it does not need to be divided into subcellular compartment (1); (2) it is with the good convergent-divergent of magnification; (3) it does not need the parameter that the user can be provided with; (4) its total intensity or Strength Changes between pair cell to image is insensitive; (5) it is based on allowing us to test interferences (for example, noise, irregularly shaped etc.) effect and seeking and stablize the model of measuring; And (6) its can be whole and/or use in unicellular level.

1.12 The optimization of parameter. the selection of best metric

If at least two points (and corresponding image) that can be used as the reference of big group of image that must be analyzed are arranged, can use at the quality metric that 1.10 parts are described.The example of this arrangement will become a plate, and described strip has some holes that are used as negative and positive control, and other hole that is used as instrument connection.In the dose curve experiment, entire curve can be used for calculated mass.In case sample and quality metric that it will be applied to are established, may cause to make parameter optimization to reach the problem of maximum possible quality.Similarly, if severally have the measuring of identical biological significance by algorithm (for example, slope 1 or slope 3; Single slope or global slopes) obtain best can on quality base, selecting in the middle of them.

Transposition described herein is measured without any real user definition parameter, at least as the ring ¹Width be in user's the same meaning of parameter.But, there are some to be established to parameter in the algorithm, it may need new cell type or specific dyeing adjustment or benefit from it.For example, the parameter control detection of mark described in 1.5 parts and watershed divide.Optimized suitable method is known in this area [13].

Optimized application in practice can be different.Positive and negative control may reside on each plate, once be used for one group of plate, or (in some cases) is calculated rather than measured.In the dose curve experiment, each curve can be optimized separately, or optimizes and can the control curve of appointment be carried out, or the like.

Embodiment 2. Film is to the easy bit test of tenuigenin

Present embodiment is described the another kind of exemplary embodiment of this theory: film is to the easy bit test of tenuigenin (or tenuigenin is to film).In this test, translocate to tenuigenin from the plasmalemma of cell such as the part of the mark of protein etc.

Figure 16 represents the image of Sports Series of the survivaling cell of GFP mark.Herein, left picture is negative status (no transposition), and midway film is intermediateness (some transposition), and right picture is positive (significantly transposition).

Figure 17 represents the signal of film in the tenuigenin transposition and the model of the joint distribution of counterstain.Herein, the top picture represents that the location of film and end picture represent cytoplasmic location.Counterstain has the high level dyeing in nuclear, and low but be not the tenuigenin dyeing of zero level.

Figure 18 represents the distribution from true cell.Herein, top and bottom field are represented film and cytoplasmic location respectively, as shown in figure 17.More than identical the measuring of Ding Yi slope can be used to describe the feature of film to cytoplasmic transposition.In addition, histogram has hinted the use of measuring, described measure according to the slope section to Y-axis prolongation on point, as shown by the rectangle of dotted line in Figure 17 and 18.

Figure 19 represents the model of the joint distribution of signal and counterstain, and the counterstain that is used for testing herein is not nucleus dyeing, but film dyeing.The location situation of described film can be described by independent slope, but two Slope Parameters are used in the tenuigenin location.

More generally, the initial histogram with 256*256 unit (bin) can be divided into more coarse grid, as shown in Figure 17.Size that can selected cell (bin) is to be provided at the observation of the fair amount in each unit (bin).Then, each 2D histogram becomes the vector in the N-dimension space, and N-is the numeral of unit (bin) herein.This permission is handled this problem and using method as pattern recognition problem ⁸All are available integrated.

Embodiment 3. The diffusion of granular reorganization test

3.1 Background

Response stimulates, and cellular component can rearrange from being diffused into granular subcellular fraction pattern (with vice versa), such as having the cell therapy of regulating son.For example, protein can add to (and/or moving to) subcellular fraction matter zone (for example vesicle) and arrive subnucleus zone (for example, PML body) to respond the treatment that is fit to part.Therefore, need system's (comprising method, algorithm and equipment) in cell, on cell or around the diffusion of cell, to change, such as adjusting that in shaker test, is exposed to a plurality of unknown effects to measure under different test conditions reporter.

3.2 Receptor activation (Transfluor ) test

Transfluor  test is (by Xsira Pharmaceuticals ^TMManage) be used to measure the activity of G-G-protein linked receptor (GPCRs).This test is adopted and is fused to the green fluorescent protein (GFP) of beta-protein inhibitor as acceptor.The basis of test is a subcellular location of measuring this fusion, and it changes according to receptor active.Especially, this fusion is according to the variation of receptor active (for example part combination) distribute from diffusion tenuigenin position to granular tenuigenin (and/or film is relevant).Because beta-protein inhibitor relates to the adjusting of many GPCRs, it is considered as general test, that is to say, one can be as the test of the activity of measuring different classes of GPCRs.

Receptor internalization in Transfluor  test causes that image presents the more granular distribution of reporter.Especially, reporter becomes less homogeneous in intracellular distribution, with " place (spots) " and " round dot (dots) " that forms concentrated reporter signal.The embodiment of Transfluor image is shown in Figure 20.Described image uses the object lens collection of the 10X magnification that has 2*2 dividing elements (binning).The diffusion profile that picture A represents not have " feminine gender " cell of GPCR activation and presents reporter." centre " cell of the GPCR that the part of representing picture B to have activates.Picture C represents to have " positive " cell of the GPCR that activates fully and presents the granular distribution of reporter.

3.3 Analyze the method for Transfluor  image

This theory is provided for analyzing the method for Transfluor  image.In certain embodiments, the method can make the intuitive concept formalization of granularity in the simple measurement.For example, the method can adopt notion known in mathematical morphology as magnitude classification [11], granulometry [15], pattern spectrum [14] and granular spectrum [17].Produce distribution by the original image series opening operation that has the structural element that increases size.In corrosion step, the value of each pixel is set to the minimum value of the environment pixel (for example, four pixels in its corner and side, or fully around eight pixels of pixel, or the like) corresponding to it.In expansion step, the value of each pixel is set to the maximal value corresponding to its environment pixel.Each opening operation can comprise one and more in succession corrosion step, and one or more expansion step are in succession followed thereafter.The size (with the size of structural element) of the quantity decision opening operation of corrosion (and expansion) step.For example, the opening operation of size " " produces by independent corrosion and expansion step, and the opening operation of size " two " produces by two corrosion steps of being followed by two expansion step, or the like.Behind each opening operation, the volume of the image that the opening operation of generation is crossed is calculated as the sum total of all pixels.

How the opening operation of Figure 21 table increase size influences and has varigrained image.Luminance Distribution is shown in the picture A-C that obtains by cell (be designated as in each picture by the line that passes cell and embed).Three pictures are from left to right represented state feminine gender, middle and positive of Transfluor  test.In the chart of each picture, present the Luminance Distribution that is used for original image (top distribution), the image (bottom surface distribution) of image (intermediate distribution) by size 1 structural element opening operation and the structural element opening operation by size 4.

Difference with the volume of the image of different opening operation size opening operation is granular spectrum, provides by following formula:

G(n)＝V(γ _n-1(X))-V(γ _n(X))

X is an image herein, and n is the opening operation size, also is referred to as thickness, and G (n) is the granular spectrum at n opening operation, γ _n(X) be n the opening operation of image X, V (X) is the volume (pixel value summation) of image X.The granular spectrum of feminine gender, centre and positive for test is shown in Figure 22, and it has the size with respect to the opening operation of drawing at the image volume percentage of this opening operation (y-axle) (x-axle).For the feature of the different conditions of describing test, we introduce and are called measuring of relative granularity, provide by following publicity:

RG＝G(T1)/G(T2)

Herein, RG is relative granularity, and T1 is the maximum characteristic thickness of granular (positive) state of test, and T2 is the maximum characteristic thickness of dispersion (feminine gender) state of test.T1 and T2 are necessary for independent value, and can be thickness ranges, adopt the mean value of granular spectral value under this situation.Using regional opening operation [16] to replace opening operation can be useful with the granular spectrum that produces.

In order to study magnification and the influence of image size to relative granularity, we use the z value, because can not obtain detailed dose curve.Two groups of images that are used to test: one group is used for positive and one group and is used for negative status.In every group, an image uses the 10X object lens to obtain, and one is used the 20X object lens to obtain, and the two all has 2 and takes advantage of 2 dividing elements (binning); Therefore according to spatial resolution, we claim that herein they are 5X and 10X magnification.Has the benefit that makes chart can in contrast to described other test like this.The image of 20X is corresponding to 1/4th in the middle of the 10X image.In addition, our image that uses be the 10X image the centre 1/4th.Each of three kinds of images is divided into four segments and negative and positive, and each segment is calculated an experimental measurement-relative granularity.Z numerical value uses feminine gender and positive setting to calculate then.Figure 23 represents the window of good test achievement (with shown in the ellipse of dotted line), at 2X and above magnification, and 0.4mm ²The image size.

More than Biao Shu algorithm can have the feature of several expectations: (1) does not need to cut apart, (2) with magnification convergent-divergent well, (3) has clearly biological significance, (4) do not need to be provided with the Any user parameter, and (5) are insensitive to total image intensity, and it can cause by the difference of camera setting.

Embodiment 4. Exemplary embodiment

Present embodiment has been described the embodiment that this theory is selected, and is expressed as the paragraph of a series of numberings:

1. computing method of measuring of the joint distribution of reporter in the biological cell, it comprises: at least two observable reporters in cell (A) are provided; (B) obtain the digital picture of reporter in the cell; And the Two dimensional Distribution at least of (C) using the reporter image value is calculated the feature of measuring of cell condition.

2. the method for paragraph 1, N reporter wherein arranged, and use therein step comprises the formation step of square figure at least always, described histogram is selected from the N-dimension histogram of the reporter value in the image sets of same object, some 2-dimension histograms of the reporter value in the image sets of same object, and the dimension of the reporter value in the image sets of same object 2 and N between the group formed of some histograms.

3. the method for paragraph 1, use therein step comprises the step of the strength specificationization (partial equilibriumization) that makes at least one reporter image.

4. the method for paragraph 1, use therein step is to carry out on independent cell base one by one.

5. the method for paragraph 1, use therein step are that the cell subclass in the image (can be independent cell or for as a whole subgroup) is carried out.

6. the method for paragraph 1, use therein step is carried out entire image, the cell that nonrecognition is independent.

7. the method for paragraph 1, use therein step comprise the step of removing pseudomorphism from image.

8. the method for paragraph 2, use therein step also comprise makes model be adapted to N-dimension histogram, and wherein said measuring is model parameter.

9. the method for paragraph 1, wherein first reporter is related with cellular compartment, and second reporter is related with protein (or other material), and described protein can change its position from a cellular compartment to another cellular compartment under experiment condition.

10. the method for paragraph 1, wherein said first reporter is related with nucleus, and second reporter related with protein (with other material), described protein can be from tenuigenin to the nucleus under experiment condition, or nucleus changes its position to tenuigenin.

11. the method for paragraph 1, wherein first reporter is related with nucleus, and second reporter is related with protein (with other material), and described protein can be from the cell membrane to tenuigenin under experiment condition, or tenuigenin changes its position to cell membrane.

12. the method for paragraph 1, wherein first reporter is related with cell membrane, and second reporter is related with protein (or other material), and described protein can be from the cell membrane to tenuigenin under experiment condition, or tenuigenin changes its position to cell membrane.

13. the method for paragraph 8 and 10, wherein said model is the straight-line segment of variable-length, described variable-length is similar to the protein reporter of transposition the right with respect to the distribution of nucleus reporter (for example, as shown in Figure 20), and wherein said measuring is the slope of this line.

14. the method for paragraph 8 and 11, wherein said model is based on the straight-line segment of variable-length, its protein reporter that is similar to transposition is with respect to (for example examining reporter, as shown in Figure 26) the right of distribution, and the wherein said statistic of measuring the subclass that is distribution (for example, as shown in Figure 26) mid point.

15. the method for paragraph 8 and 12, wherein said model is based on the straight-line segment of variable-length, and it is similar to the protein reporter of transposition the right with respect to the distribution of film reporter (for example, as shown in Figure 28).

16. the method for paragraph 2, wherein said N-dimension histogram is regarded as M-dimensional vector (M is the sum of the unit (bin) in such histogram), wherein each cell (or cell mass, or entire image) is regarded as the point in the M-dimension space, wherein uses the mode identification method analysis of cells.

17. the method for paragraph 16, wherein such mode identification method are the cytological classification methods in the predetermine class, wherein said measuring is degree and the class name similar to such classification.

18. the method for paragraph 1, wherein the reporter image is pretreated for making (for example, because kernel and the filler opening) reduction of some features of not expecting or it is proofreaied and correct, or strengthens the feature of some expectations.

19. the method for paragraph 4, cell mass are measured by the statistics of position or measure and describe feature by the statistics of variation.

20. the method for paragraph 19, measuring of wherein said position is selected from the group of being made up of mean value, intermediate value, mode etc.; And the group of measuring the compositions such as intermediate value deviation that are selected from standard deviation, center on intermediate value of wherein said variation.

21. the method for paragraph 4, wherein cell mass is described feature by the unicellular histogrammic principal component analysis (PCA) (PCA) of measuring distribution.

22. the method for paragraph 1, wherein measuring is that the name (classification) of cell state is measured, for example the stage in cell cycle cycle.

23. the method for paragraph 1, the step that wherein obtains digital picture is carried out simultaneously at least two different reporters.

24. the method for paragraph 1, the step that wherein obtains digital picture is carried out in succession at least two different reporters.

25. the method for paragraph 1, wherein said tolerance be at low magnification, for example≤and 2X object lens (～〉=5 μ m/ pixel).

26. the computing method of measuring of the joint distribution of reporter in the biological cell, it comprises: at least two observable reporters in cell (A) are provided; (B) under at least two test conditions, obtain the digital picture of reporter in the cell; (C) Two dimensional Distribution at least of use reporter value is calculated the tolerance feature of cell condition; And (D) under at least two test conditions, be provided at cell and measure the quality metric that calculates.

27. the method for paragraph 26, use therein step comprise the image analysis method that relies on the digital parameter of a cover.

28. the method for paragraph 27, the value of wherein selecting digital parameters is to optimize the quality metric that calculates on cell is measured under at least two test conditions.

29. the method for paragraph 26, use therein step comprise at least two kinds of selections of calculating the method for cells tolerance and providing the method that best in quality measures under at least two test conditions.

30. the method for paragraph 26, use therein step comprises the step of selecting image subset, described subclass provides best in quality and measures (for example, the best region-great majority of camera view of the best of the system in the hole or the system in camera view be used to focus on reason).

31. the method for paragraph 28,29 and 30, wherein said selection (optimization) put in one of at least two test conditions and carry out and be applied to other test condition.

32. the method for paragraph 26, wherein said test condition are the variable concentrations of reagent.

33. the method for paragraph 32, wherein said reagent are candidate's medical compoundss.

34. the method for paragraph 26, wherein said test condition are the different time points of a certain process.

35. will have the image segmentation of biological cell is the method that contains the segment of unicellular or cell mass, it is included in and carries out watershed transform on the image, and described image is the combination of at least two reporter images.

Above-describedly openly can comprise a plurality of distinct invention that has separate utility.Even each of these inventions is open with its preferred form, be not considered to the meaning that limits as its specific implementations disclosed and explanation in this article, because many variations are possible.Subject matter of an invention comprise all of different element disclosed herein, feature, function and/or characteristic new with non-obvious combination and sub-portfolio.Following claim particularly points out and is considered to new and non-obvious some combination and sub-portfolio.The invention of implementing with the combination of feature, function, element and/or characteristic and sub-portfolio requires power in can applying for, described application has required the preference of this and related application.Such claim, no matter whether point to different inventions or identical invention, and whether wideer, narrower, be equal to, or be different from the scope that original rights requires, also be believed to comprise within theme of the present invention of the present disclosure.

Reference

1. U.S. Patent number5,989,835, title " is used for the system based on the screening of cell.”

2.PCT publication No. WO 03/078965, title " is used for that auto color is cut apart and for the system and method for the minimum significant response of the part local strength tolerance of cellular compartment.”

3. U.S.'s publication No. 2003/0059093, and title " is used for determining the method for cells of interest compartment tissue.”

4. U.S.'s publication No. 2003/0202689, and title " is used for the graphical analysis based on radiation of biological sample.”

5.S.Beucher with " morphological method of cutting apart: watershed transform " of F.Meyer, the morphology in the Flame Image Process, E.R.Dougherty-Ed., pp.433-481, Marcel Dekker, New York, 1993..

6.L.Vincent with " watershed divide in the digital space: " of P.Soille, IEEE Transactions of Pattem Analysis and Machine Intelligence, 13, No.6, pp.583-598,1991 based on the efficient algorithm that immerses simulation.

7.I.Ravkin, V.Temov, A.Nelson, M.Zarowitz, M.Hoopes, Y.Verhovsky, G.Ascue, S.Goldbard, O.Beske, B.Bhagwat and H.Marciniak " using the multichannel high flux image blood count of code carrier ", Proc.SPIE Vol.5322, pp.52-63,2004 (imaging of biomolecule, cell and tissue, processing and analysis II; D.Nicolau, J.Enderlein, R.Leif and D.Farkas; Eds.)

8." pattern classification " of R.Duda, P.Hart and D.Stork, Wiley-Interscience, 2nd Ed., 2000.

9.J.Zhang, " the simple statistics parameters " J.Biomol.Screening 4:pp.67-73 that is used for the calculating and the affirmation of high throughput screening assay, 1999 of T.Chung and K Oldenburg.

10. Flame Image Process kit, The MathWorks, Inc.

http://www.mathworks.com/products/image/。

11 ." graphical analysis and the mathematical morphology " of J.Serra, Vol.1.Academic Press, London, 1989.

12.I.Jolliffe " principal component analysis (PCA) ", Springer-Verlag, 2nd Ed., 2002.

13.R.Fletcher " practical approach of optimization ", Wiley, 2nd Ed., 2000.

14.P.Maragos " pattern spectrum and multiple dimensioned shape are represented ", IEEE Transactions onPattern Analysis and Machine Intelligence, 11, N7, pp.701-716,1989.

15.L.Vincent " granulometry and opening operation tree ", Fundamenta Informaticae, 41, No.1-2, pp.57-90, IOS Press, 2000.

16.L.Vincent " being used for the morphology area opening operation and the closed operation of grayscale image ", Proc.NATO Shape in Picture Workshop, Driebergen, The Netherlands, pp.197-208,1992.

17.I.Ravkin with " bit representation technology and the Flame Image Process " of V.Temov, Applied Informatics, v.14, pp.41-90, Finances and Statistics, Moskow, 1988 (in Russian).

1.U.S.Patent?No.5,989,835，entitled″System?for?cell-based?screening.″

2.PCT?Publication?No.WO?03/078965，entitled″System?and?method?forautomatic?color?segmentation?and?minimum?significant?response?formeasurement?of?fractional?localized?intensity?of?cellular?compartments.″

3.U.S.Publication?No.2003/0059093，entitled″Methods?for?determining?theorganization?of?a?cellular?component?of?interest.″

4.U.S.Publication?No.2003/0202689，entitled″Ray-based?image?analysis?forbiological?specimens.″

5.S.Beucher?and?F.Meyer，″The?Morphological?Approach?to?Segmentation：The?Watershed Transformation″in：Mathematical?Morphology?in?ImageProcessing，E.R.Dougherty-Ed.，pp.433-481，Marcel?Dekker，New?York，1993.

6.L.Vincent?and?P.Soille，″Watersheds?in?Digital?Spaces：An?EfficientAlgorithm?Based?on?Immersion?Simulations″，IEEE?Transactions?of?PattemAnalysis?and?Machine?Intelligence，13，No.6，pp.583-598，1991.

7.I.Ravkin，V.Temov，A.Nelson，M.Zarowitz，M.Hoopes，Y.Verhovsky，G.Ascue，S.Goldbard，O.Beske，B.Bhagwat，and?H.Marciniak，″Multiplexedhigh-throughput?image?cytometry?using?encoded?carriers″，Proc.SPIE?Vol.5322，pp.52-63，2004(Imaging，Manipulation，and?Analysis?of?Biomolecules，Cells，and?Tissues?II；D.Nicolau，J.Enderlein，R.Leif，and?D.Farkas；Eds.)

8.R.Duda，P.Hart，and?D.Stork，Pattern?Classification，Wiley-Interscience，2nd?Ed.，2000.

9.J.Zhang，T.Chung，and?K?Oldenburg，″A?simple?statistical?parameter?for?usein?evaluation?and?validation?of?high?throughput?screening?assays，″J.Biomol.Screening?4：pp.67-73，1999.

10.Image?Processing?Toolbox，The?MathWorks，Inc.http://www.mathworks.com/products/image/.

11.J.Serra，Image?Analysis?and?Mathematical?Morphology，Vol.1.AcademicPress，London，1989.

12.I.Jolliffe，Principal?Component?Analysis，Springer-Verlag，2nd?Ed.，2002.

13.R.Fletcher，Practical?Methods?of?Optimization，Wiley，2nd?Ed.，2000.

14.P.Maragos，″Pattern?spectrum?and?multiscale?shape?representation″，IEEETransactions?on?Pattern?Analysis?and?Machine?Intelligence，11，N?7，pp.701-716，1989.

15.L.Vincent，″Granulometries?and?Opening?Trees，″FundamentaInformaticae，41，No.1-2，pp.57-90，IOS?Press，2000.

16.L.Vincent，″Morphological?Area?Opening?and?Closing?for?GrayscaleImages，″Proc.NATO?Shape?in?Picture?Workshop，Driebergen，The?Netherlands，pp.197-208，1992.

17.I.Ravkin?and?V.Temov，″Bit?representation?techniques?and?imageprocessing，″Applied?Informatics，v.14，pp.41-90，Finances?and?Statistics，Moskow，1988(in?Russian).

Claims (35)

1. one kind is calculated the method that the reporter joint distribution is measured in the biological cell, and it comprises:

At least two reporters are provided, and it is visual in cell;

Obtain the digital picture of reporter described in the cell; And

The Two dimensional Distribution at least of described image value of using reporter is with the feature of measuring of the condition of calculating described cell.

2. the method for claim 1, N reporter wherein arranged, and the step of wherein said use comprises the formation step of square figure at least always, described histogram is selected from the N-dimension histogram of the reporter value in the image sets of same object, a plurality of 2-dimension histograms of the reporter value in the image sets of same object, and the dimension of the reporter value in the image sets of same object 2 and N between the group formed of a plurality of histograms.

3. the method for claim 1, the step of wherein said use comprises the step of the strength criterionization that makes at least one described reporter image.

4. the method for claim 1, the step of wherein said use is carried out on the basis of independent cell one by one.

5. the method for claim 1, the step of wherein said use is carried out the cell subclass in the described image.

6. the method for claim 1, the carrying out of the step of wherein said use do not have the identification of separate cell.

7. the method for claim 1, the step of wherein said use comprise the step of removing pseudomorphism from described image.

8. method as claimed in claim 2, the step of wherein said use also comprise makes model be suitable for described N-dimension histogram, and wherein said measuring is the parameter of described model.

9. the method for claim 1, wherein said first reporter is related with cellular compartment, and described second reporter is related with a protein, and described protein can change its position from a cellular compartment to another cellular compartment under experiment condition.

10. the method for claim 1, wherein said first reporter is related with described nucleus, and described second reporter is related with a protein, described protein can be from tenuigenin to the nucleus under experiment condition or nucleus change its position to tenuigenin.

11. the method for claim 1, wherein said first reporter is related with described nucleus, and described second reporter is related with a protein, described protein can be from the cell membrane to tenuigenin under experiment condition or tenuigenin change its position to cell membrane.

12. the method for claim 1, wherein said first reporter is related with described cell membrane, and described second reporter is related with a protein, described protein can be from the cell membrane to tenuigenin under experiment condition or tenuigenin change its position to cell membrane.

13. method as claimed in claim 10, wherein there be N reporter, the step of wherein said use comprises the formation step of square figure at least always, described histogram is selected from the N-dimension histogram of the reporter value in the image sets of same object, a plurality of 2-dimension histograms of the reporter value in the image sets of same object, and the dimension of the reporter value in the image sets of same object 2 and N between the group formed of a plurality of histograms, wherein said step also comprises makes model be suitable for described N-dimension histogram, wherein said measuring is the parameter of described model, wherein said model is the straight-line segment of variable-length, the right that its protein reporter that is similar to described transposition distributes with respect to the nucleus reporter, and wherein said measuring is the slope of this line.

14. method as claimed in claim 11, wherein there be N reporter, the step of wherein said use comprises the formation step of square figure at least always, described histogram is selected from the N-dimension histogram of the reporter value in the image sets of same object, a plurality of 2-dimension histograms of the reporter value in the image sets of same object, and the dimension of the reporter value in the image sets of same object 2 and N between the group formed of a plurality of histograms, wherein said step also comprises makes model be suitable for described N-dimension histogram, wherein said measuring is the parameter of described model, wherein said model is based on the straight-line segment of variable-length, the right that its protein reporter that is similar to described transposition distributes with respect to the nucleus reporter, the wherein said statistic of measuring the subclass that is described distribution mid point.

15. method as claimed in claim 12, wherein there be N reporter, the step of wherein said use comprises the formation step of square figure at least always, described histogram is selected from the N-dimension histogram of the reporter value in the image sets of same object, a plurality of 2-dimension histograms of the reporter value in the image sets of same object, and the dimension of the reporter value in the image sets of same object 2 and N between the group formed of a plurality of histograms, wherein said step also comprises makes model be suitable for described N-dimension histogram, wherein said measuring is the parameter of described model, wherein said model is based on the straight-line segment of variable-length, and it is similar to protein reporter of described transposition the right with respect to the distribution of film reporter.

16. method as claimed in claim 2, wherein said N-dimension histogram is regarded as M-dimensional vector (M is the described sum of the unit in this histogram), wherein each cell (or cell mass or described entire image) is regarded as the point in described M-dimension space, and wherein uses the mode identification method analysis of cells.

17. method as claimed in claim 16, wherein this mode identification method be cytological classification in predetermine class, wherein said measuring is relatively this type of other similarity and class name.

18. the method for claim 1, wherein the reporter image is pretreated with reduction or proofread and correct the feature that some are not expected, or strengthens the feature of some expectations.

19. method as claimed in claim 4, wherein said cell mass are measured by the statistics of position or measure demarcation by the statistics of variation.

20. it is to be selected from the group that mean value, intermediate value and mode are formed that method as claimed in claim 19, wherein said position are measured, and measuring of wherein said variation is to be selected to comprise standard deviation and the group that centers on the intermediate value deviation of intermediate value.

21. method as claimed in claim 4, wherein said cell mass is demarcated by the described unicellular described histogrammic principal component analysis (PCA) (PCA) of measuring distribution.

22. the method for claim 1, wherein measuring is that the name (classification) of cell state is measured.

23. the method for claim 1, the step of wherein said acquisition digital picture is carried out simultaneously at least two different reporters.

24. the method for claim 1, the wherein said step that obtains digital picture is carried out in succession at least two different report bases.

25. the method for claim 1, wherein said measuring is at low magnification.

26. a method of measuring of calculating reporter joint distribution in the biological cell, it comprises:

At least two reporters are provided, and it is visual in cell;

Under at least two test conditions, obtain the digital picture of reporter described in the cell;

At least the Two dimensional Distribution of use reporter value is calculated the tolerance feature of described cell condition; And

Under described two test conditions, be provided at cell at least and measure the quality metric that calculates.

27. method as claimed in claim 26, the step of wherein said use comprise the image analysis method according to the digital parameter of a cover.

28. method as claimed in claim 27, the value of wherein selecting described digital parameters at least under described two test conditions is to optimize the described quality metric that calculates on cell is measured.

29. method as claimed in claim 26, the step of wherein said use comprise that at least two kinds are calculated the method that cell is measured, and the selection that provides the described method of best quality metric at least under described two test conditions.

30. method as claimed in claim 26, the step of wherein said use comprise that selection provides the step of the image subset of best quality metric.

31., wherein saidly be selected in one under at least two kinds of test conditions and put and carry out and be applied to other test condition as claim 28,29 and 30 each described methods.

32. method as claimed in claim 26, wherein said test condition are the variable concentrations of reagent.

33. method as claimed in claim 32, wherein said reagent is candidate drug compounds.

34. method as claimed in claim 26, wherein said test condition are the different time points of a certain process.

35. the image segmentation that will have biological cell is the method that contains the segment of unicellular or cell mass, it is included in and carries out watershed transform on the image, and described image is the combination of the image of at least two reporters.

Images