版權(quán)說(shuō)明:本文檔由用戶(hù)提供并上傳,收益歸屬內(nèi)容提供方,若內(nèi)容存在侵權(quán),請(qǐng)進(jìn)行舉報(bào)或認(rèn)領(lǐng)
文檔簡(jiǎn)介
1、Flow Cytometry Ming Yan2, Joe Trotter2, Rudi Varro2, Francis Mandy1, Diether Recktenwald21 Soft Flow Hungary Ltd., H-7628 Kedves 20 u. Pecs, Hungary2 BD Biosciences, 2350 Qume Drive, San Jose, CA 95131-1812Flow cytometry allows the analysis and sorting of particles of biological interest at rates of
2、 more than 104 s-1, based on the analysis of light scatter and fluorescence. It also permits the quantitative analysis of many cellular constituents based on fluorescence measurements. Measurements at the single-molecule level have been reported. Based on the versatility and richness of information
3、of flow cytometry, it is used in biological and biomedical research, and for clinical data collection. This chapter provides an overview of flow cytometry and its biomedical and clinical applications. For readers interested in further details, we provide references to additional reviews of subtopics
4、. A comprehensive account of flow cytometry, covering all aspects until 1995, can be found in Practical Flow Cytometry (Shapiro 1995).HardwareFlow cytometers measure multiple optical properties of particles generally without spatial resolution from about 20 m down to submicroscopic size at rates of
5、several thousand per second. A typical flow cytometer consists of a fluidic system, the optical components, and analog and/or digital electronics for data processing, storage, and evaluation (Fig 1). Special cytometers also sort particles into different fractions, based on optical particle propertie
6、s. Several sorting mechanisms have been used as described below.FluidicsIn the fluidics system of a cytometer, a particle suspension from a tube or well of a microtiter plate is injected into a second fluid stream of aqueous sheath fluid - mostly saline - to create a very narrow, quasi one-dimension
7、al file of particles for intersection with a light beam for optical measurements. Typical stream velocities are 10 m/s. The ratio of sheath fluid volume to sample volume with the size of the observation cuvette determines the diameter of the sample stream, and influences the precision of the optical
8、 measurements. Typical total stream diameters are on the order of 100um, sample streams on the order of 10um. The sample stream diameter can be controlled by the differential pressure between the sheath and sample fluid (Peters et al. 1985). The sample stream diameter usually relates to the sample f
9、low rate. The low flow rate corresponds to the small sample stream diameter. Coefficients of variation (CV) better than 2% are quite common for measurements of the DNA content of cell nuclei. Typically, the better CV is achieved by low sample flow rate due to uniform light beam illumination on the e
10、ntire sample core stream. Optics To perform optical measurements on the particle stream in a cytometer, a light beam, commonly from a laser, is focused on the center of the fluid stream either in a cuvette or a free flowing stream in air for most cell sorters. Fig. 2 shows a typical optics diagram f
11、or a three laser sorter. Mercury arc lamps and LEDs are also used as excitation light sources. Cylindrical lenses are used frequently to achieve a better uniformity of light intensity for the particle illumination with a typical beam height of about 10-20um. Research flow cytometers are capable of m
12、easuring several light scatter and fluorescence emission intensities. Many fluorescence measurements on biological systems require a very low limit of detection. Therefore high numerical aperture lenses are used for the collection of emitted light. Dichroic filters separate scattered and fluorescent
13、 light into separate wavelengths bands. Photomultipliers (PMTs) measure the light intensities in the different wavelength bands. With an optimized instrument, photon-noise-limited measurements can be performed, and less than 200 molecules per particle (approx. 10-21 moles) of some fluorescent dyes c
14、an be detected. Some newer specialized systems use avalanche photodiodes in place of PMTs; light scatter can be detected with photodiodes. The number of detectors has increased significantly due to multiple color assay development. The numerous designs of detector arrays have been deployed for bette
15、r compact design and low loss performance. Figure 3 shows fiber optic linked PMT detector array design (Oostman et al. 2006). High-end research systems offer multiple excitation light sources, either in a co-linear arrangement or for more flexibility for the resolution of multiple fluorophores with
16、completely separate spatial optical paths and temporal synchronization.ElectronicsSignals from the light detectors are amplified. To allow the measurement of small and large signal intensity ranges linear and logarithmic (typically 4 decades) ranges are provided on most instruments. Signal subtracti
17、on circuitry, linking adjacent spectral fluorescence emission bands, allow for the correction for spectral overlap between fluorescent dyes to express the intensity measurements in relative units of dye concentration rather than light intensities. After baseline subtraction, an analog pulse height o
18、r pulse area, and a calculated pulse width are provided to an analog-to-digital converter (ADC). The ADC is triggered by a pulse height threshold, based on a boolean combination of the measurement parameters, and after digitization all of the particle measurements including those from separate light
19、 beams - after temporal synchronization - are stored as a record for the particle in the data matrix. Approaches for the extraction of population information from this data matrix are described in the Data Analysis section below.Recently, digital electronics has been used to derive the particle meas
20、urements from a continuous digitization of detector output through high speed ADCs. All of the signal calculations on the pulses, including height, area, width, logarithms, and spectral overlap correction are performed with high speed digital signal processors. Calculations of signal parameters are
21、performed with higher accuracy than by analog approximation, and the approach also provides more flexibility for future applications of modern signal signature analysis for flow cytometry.Cell sorting with cloningAs mentioned above, several approaches have been used to sort particles in essentially
22、real time, based on the measurements of a flow cytometer. Cell sorting has been reviewed in detail in a recent book chapter (Hoffman & Houck 1998). The most common method for flow cytometric cell sorting uses a piezoelectric actuator, which breaks the stream containing the particles under analysis i
23、nto droplets. Charging the droplets at the right time with a charge pulse, triggered by the results of the optical analysis allows the electrostatic deflection into typically four positions, where a collection tube can be placed. Sorting rates of higher than 104 per second can be achieved for the pu
24、rification of millions of purified particles (cells) with specific properties in under an hour. In another setup the particles of interest are deflected into a multi-well plate (typically 96 wells in an 8x12 well arrangement) or onto a plate with cell nutrient under xy control. For cloning, single c
25、ells can be deposited there to achieve a cultured cell population, which can be traced back to an individual cell. (Fig.3).Data AnalysisAs mentioned above, data from flow cytometric measurements are stored in a data matrix, where each row contains the measurement for an individual particle with a co
26、lumn for each of the measurement parameters. One of the measurement parameters can be a time tag for kinetic data evaluation. For most analyses information on at least 5000 particles is stored, in many cases substantially more. At least one byte of data is stored for each of the parameters; in many
27、cases the resolution is higher. Newer digital instruments store data as high resolution floating point numbers. This kind of data file is called a listmode file. This designation comes from the times when data storage was very expensive. To economize on usage of memory, data was also stored as histo
28、gram data (see below).The listmode data matrix is the basis for all subsequent data analysis, but by itself cannot be evaluated easily without computer based data transformation.A simple analysis of the data consists of calculating a histogram of the number of particles at each parameter value for a
29、ll of the parameters. In the most simple implementation the digital parameter value for each of the particles in the matrix is used as an address for an array (1024 elements long for a 10 bit parameter resolution), and the content of the corresponding array element is incremented by 1, every time th
30、e respective value is observed. Data histograms (Fig. 4a) provide a view of intensity distribution for the subpopulations in a particle ensemble; however, information about the correlation between parameters is lost.A so-called dotplot (Fig. 4b) shows the intensity distribution with pairs of two mea
31、surement parameters. In a two-dimensional plot dots are displayed for each particle at x y coordinates corresponding to two parameter values. This plot shows population locations and widths with some relative frequency information from the dot density. Quantitative frequency information is lost, bec
32、ause overlapping dots show only as one. Density and contour plots show quantitative population frequency information with two parameters. Both are based on two-dimensional histograms. A density plot uses a gray scale or colors to represent the z-axis (frequency information), whereas a contour plot u
33、ses different contour levels to show histogram height or probability contours show population frequency.All the data representations above show only up to two parameters correlated simultaneously. The listmode data matrix may contain many parameters, for most measurements at least four. Therefore th
34、e problem remains to look at subpopulations of the particle ensemble in multi-dimensional space. A process called gating uses one or two-dimensional dotplot displays to select parameter value sets to include or exclude from additional data displays. In this way properties of particle subpopulations
35、can be determined for all measured parameters. A special case of novel multi-parameter gating is detailed in (Bierre & Thiel 1998). Several software packages offer additional features for multi-parameter data analysis. Paint-a-GateTM (Conrad, Reichert & Bezdek 1989) uses color to highlight populatio
36、ns selected in one display in five more two-dimensional displays in real-time. For automated multi-parameter analysis density-based or nearest neighbor cluster algorithms have been used amongst others. Further details on data analysis methods for flow cytometry have been discussed elsewhere (Watson
37、1992) Flow Cytometry MeasurementsA flow cytometric measurement characterizes one or more populations of a particle suspension with a count value and several optical parameter intensities and their distributions. From the partial count, particle concentrations can be calculated, if the sample volume
38、is known. Regardless of the capabilities of the instrumentation to record sample volumes, a known concentration of a reference particle with scatter and fluorescence properties different from the particles of the sample can be added to obtain concentrations of all of the particle populations. Howeve
39、r, many applications of flow cytometry only report the relative frequency of a sample particle subset as fraction of subset in a superset of the particle sample i.e. T-cells as a percentage of lymphocytes.The optical parameter intensities are used to derive physical properties of the particles in a
40、sample. Light scatter intensities are related to the size and index of refraction of particles; fluorescence intensities are related to the mass of fluorophors per particle. With calibration of the system absolute masses of analyte per particle can be determined from fluorescence intensities. Howeve
41、r, for most application relative quantities are reported. Limits of detection for fluorescence measurements with commercial systems are on the order of a few hundred fluorescent molecules per particle (Coventry et al. 1994); with specialized instrumentation single-molecule detection has been achieve
42、d (Harding & Keller 1992). Flow Cytometer Characterization and SetupMeasurement of instrument characteristics and its correlation to assay performance provides quantitative criteria for quality assurance and quality control. One simple measure of sensitivity is the difference in fluorescence intensi
43、ty of a specific positive cell and unstained population divided by twice the standard deviation of the unstained population. This normalized signal-to-background approach is best used to either compare sensitivity for a specific reagent among cytometers, or to determine which dye color might be opti
44、mal for any specific marker when planning multicolor reagent panels. However, it does not directly address the underlying contributions that ultimately determine the fluorescence sensitivity of the cytometer itself, i.e., optics, fluidics, and electronics. In a multicolor assay, the ability to accur
45、ately detect spectral spillover from other fluorescence dyes is also an important factor. The cytometer performance can be simplified as instrument performance factors Q (detection efficiency), B (background), standard deviation of electronics noise (SDen), and linearity (Chase & Hoffman 1998; Gauch
46、er, Grumwald & Frelat 1988). The dye embedded polystylene bead exhibits broad fluorescence spectrum which enables us to measure the performance from a cytometer for at least 18 different colors commonly used simultaneously. A bead set consisting of different fluorescence intensity levels allows user
47、s to assess the cytometer performance including laser alignment, Q, B, SDen, and linearity within the dynamic range typically used for immunofluorescence. The instrument performance Q and B are usually normalized to a fluorescence intensity standard. The MESF is a commonly used standard (Henderson e
48、t al. 1998; Lenkei et al. 1998) with only a few fluorochrome available. Equivalent reference fluorescence (ERF) was proposed as an alternative (Wang et al. 2008) method to cross calibrate various fluorescence standards, usually traceable from lot to lot by bead manufacturer. Large fluorescence inten
49、sity variation of different antibody linked fluorochromes requires the assay specific instrument setup. Typically, a fluorochrome labeled bead or dye-embedded bead is used for instrument setup prior to running biological samples. Proper setup provides the consistency for the assay from sample to sam
50、ple. The fluorescence crosstalk from the detectors measured from multiple fluorochromes, called spillover, is usually required to know before performing an assay. The spillover is due to fluorescence spectral overlap and can be compensated mathematically during data processing. Measured fluorescence
51、 intensity distributions can be discomposited into the combination of Q, B, and electronics noise in a multicolor cytometer. Along with the background light from the spillover of other fluorescence channels, the experimental results from a multicolor assay on an instrument with known and varied comb
52、inations of these factors are compared to theoretical predictions (Hoffman & Woods 2007). The ability to predict assay performance based on measured instrument characteristics is the critical factor for quality assurance of cytometer based assay. New Approaches to Data Transformation and AnalysisDat
53、a TransformationIn flow cytometry the need often arises to measure signal changes over a significantly large dynamic range. Immunologists measure a wide range of expression levels for specific cell surface markers, and this has for many years largely been solved by employing a simple logarithmic tra
54、nsformation to span the dynamic range requirements, either by the use of a logarithmic amplifier in hardware or by displaying high resolution linear data on a log scale. In many cases the simultaneous display of a negative cell population (essentially unstained) and a positive (stained) population o
55、n the same scale necessitates two to four or more decades dynamic range, only the standard logarithmic transformation does a poor job of handling populations with low medians and high variances, and cannot handle data values less than or equal to zero. Several alternatives to the logarithmic transfo
56、rmation have been proposed and implemented in software that mitigate issues intrinsic to the log scale by allowing the lower end of the scale to be nearly linear, yet become and remain logarithmic for the upper decades, and allow for the display of values less than or equal to zero (Bagwell 2005; Pa
57、rks, Roederer & Moore 2006; Trotter 2007). After immunofluorescence data is properly compensated, and various positive population medians are translocated to be orthogonal to unstained populations in other dye dimensions, their higher variance due to photoelectron statistics results in what most inv
58、estigators term “spread” in the data. As a result, properly compensated data sets frequently contain populations with large variances and low medians, and the extents of the data often extend far below zero. Since zero and negative values remain undefined within the log transformation, that approach
59、 only satisfies the dynamic range requirement for immunofluorescence data visualization and seriously confounds the proper display of dimly stained and unstained cell populations. Figure 5 shows an example data set using both the log and Logicle transformations to demonstrate the usefulness of an ot
60、her than log transformation for immunofluorescence data that spans several decades dynamic range.AnalysisPromising new approaches have been proposed for the analysis and display of high dimensional flow data to satisfy the need to simultaneously deconvolve and display events in k dimensions. The tra
溫馨提示
- 1. 本站所有資源如無(wú)特殊說(shuō)明,都需要本地電腦安裝OFFICE2007和PDF閱讀器。圖紙軟件為CAD,CAXA,PROE,UG,SolidWorks等.壓縮文件請(qǐng)下載最新的WinRAR軟件解壓。
- 2. 本站的文檔不包含任何第三方提供的附件圖紙等,如果需要附件,請(qǐng)聯(lián)系上傳者。文件的所有權(quán)益歸上傳用戶(hù)所有。
- 3. 本站RAR壓縮包中若帶圖紙,網(wǎng)頁(yè)內(nèi)容里面會(huì)有圖紙預(yù)覽,若沒(méi)有圖紙預(yù)覽就沒(méi)有圖紙。
- 4. 未經(jīng)權(quán)益所有人同意不得將文件中的內(nèi)容挪作商業(yè)或盈利用途。
- 5. 人人文庫(kù)網(wǎng)僅提供信息存儲(chǔ)空間,僅對(duì)用戶(hù)上傳內(nèi)容的表現(xiàn)方式做保護(hù)處理,對(duì)用戶(hù)上傳分享的文檔內(nèi)容本身不做任何修改或編輯,并不能對(duì)任何下載內(nèi)容負(fù)責(zé)。
- 6. 下載文件中如有侵權(quán)或不適當(dāng)內(nèi)容,請(qǐng)與我們聯(lián)系,我們立即糾正。
- 7. 本站不保證下載資源的準(zhǔn)確性、安全性和完整性, 同時(shí)也不承擔(dān)用戶(hù)因使用這些下載資源對(duì)自己和他人造成任何形式的傷害或損失。
最新文檔
- 小學(xué)易讀錯(cuò)的漢字
- 滬科版八年級(jí)數(shù)學(xué)上冊(cè)第14章全等三角形14-2三角形全等的判定第4課時(shí)三角形全等的判定“角角邊”課件
- 人教版七年級(jí)下期語(yǔ)法聚焦句型復(fù)習(xí)提綱
- 蘇教版八年級(jí)生物上冊(cè)第7單元第二十章生物圈是最大的生態(tài)系統(tǒng)素養(yǎng)綜合檢測(cè)課件
- 企業(yè)倫理學(xué)(原書(shū)第5版) 課件 第5、6章 企業(yè)社會(huì)責(zé)任;倫理決策:雇主的義務(wù)和雇員的權(quán)利
- 2024-2025學(xué)年版塊8 運(yùn)動(dòng)和力 專(zhuān)題8-4 力的合成 (含答案) 初中物理尖子生自主招生培優(yōu)講義83講
- 大學(xué)生創(chuàng)新創(chuàng)業(yè)項(xiàng)目之DIY陶瓷作坊
- 重慶市九龍坡區(qū)2024年中考語(yǔ)文適應(yīng)性考試試卷(含答案)
- 內(nèi)蒙古呼倫貝爾市2024年中考數(shù)學(xué)全真模擬試卷含解析
- 江蘇省常州市正衡中學(xué)2024-2025學(xué)年八年級(jí)上學(xué)期期中模擬英語(yǔ)試卷
- MOOC 跨文化交際通識(shí)通論-揚(yáng)州大學(xué) 中國(guó)大學(xué)慕課答案
- EDA實(shí)驗(yàn)報(bào)告1組合邏輯電路的設(shè)計(jì)
- 科學(xué)教學(xué)設(shè)計(jì)《食物鏈》
- 力樸素抗癌藥理藥效及其合成
- 球磨機(jī)設(shè)計(jì)說(shuō)明書(shū)
- 餐飲美學(xué)餐飲空間格局設(shè)計(jì)
- 有效初三英語(yǔ)課堂教學(xué)ppt課件
- 幸福在哪里作文800字高中范文
- 五人制足球比賽記錄表.doc
- 整式的乘法與因式分解所有知識(shí)點(diǎn)總結(jié)
- 《運(yùn)動(dòng)生理學(xué)》教案
評(píng)論
0/150
提交評(píng)論