Analyzing multivariate flow cytometric data in aquatic sciences.

Flow cytometry has recently been introduced in aquatic ecology. Its unique feature is to measure several optical characteristics simultaneously on a large number of cells. Until now, these data have generally been analyzed in simple ways, e.g., frequency histograms and bivariate scatter diagrams, so that the multivariate potential of the data has not been fully exploited. This paper presents a way of answering ecologically meaningful questions, using the multivariate characteristics of the data. In order to do so, the multivariate data are reduced to a small number of classes by clustering, which reduces the data to a categorical variable. Multivariate pairwise comparisons can then be performed among samples using these new data vectors. The test case presented in the paper forms a time series of observations from which the new method enables us to study on the temporal evolution of cell types.     

A proposal for a flow cytometric data file standard

The increasing complexity of multiparameter data collection and analysis in flow cytometry and the development of relatively inexpensive arc-lamp-based flow cytometers, which increases the probability that laboratories or institutions may have more than one type of instrument, creates a need for shareable analysis programs and for the transport of flow cytometric data files within an installation or from one institution to another. To address this need, we propose a standard file format to be used for all flow cytometric data. The general principles of this proposal are: (1) The data file will contain a minimum of three segments, TEXT, DATA, and ANALYSIS; (2) The TEXT and ANALYSIS segments consist of KEYWORDS, which are the names of data fields, and their values; (3) All TEXT is encoded in ASCII; (4) KEYWORDS and their values may be of any length; (5) Certain KEYWORDS will be standard, i.e., having specified formats to be recognized by all programs. The structure of the DATA segment will be uniquely defined by the values of KEYWORDS in the TEXT area. It may be in any bit resolution, facilitating compatibility between machines with different word length and/or allowing bit compression of the data. The structured nature of the TEXT area should facilitate management of flow cytometric data using existing data base management systems. The proposed file format has been implemented on VAX, PDP-11, and HP9920 based flow cytometry data acquisition systems.     

Analysis of flow cytometric data of irradiated cell populations

Summary Flow cytometry (FCM) and autoradiography have been applied to determine changes in the cell kinetics of irradiated cells. Synchronized L-929 cells were irradiated with 10 Gy of X-rays when progressing from G₁-to S-phase of the cell cycle. In this study three methods to analyse DNA histograms were tested for applicability on FCM data obtained from cell populations blocked or retarded in the cycle: A) the Gaussian integral method, B) the peak-half-reflection method, and C) the rectangle method. Since histograms from synchronized cells are heavily distorted as compared to those obtained from exponentially growing cells and are quite similar to histograms from irradiated cells, they were used to test the suitability of the evaluation methods. Comparing the evaluated FCM data with the autoradiographic results from the same experimental series, the Gaussian integral method proved to be superior to the two other relatively simple approximation methods. The FCM histograms of irradiated cells were therefore analyzed only by the Gaussian integral method. It was shown that a considerable fraction of cells is still in the S-phase 25 h post irradiation, the DNA synthesis of which has ceased, as shown by autoradiography. This indicated that parallel measurements using FCM and autoradiography yield additional information on cell kinetic changes that cannot be obtained from applying one of the two methods used.     

New method for the analysis of flow cytometric data

A numerical method for deriving the fractions of cells in different phases of the cell cycle from a single observed DNA histogram is presented. The observed histogram is regarded as a polluted version (containing allocation errors) of the true histogram. A mathematical model is used to describe the pollution process. A theoretical histogram, representing the true histogram, is constructed so that G1 cells are put into one channel and G2M cells into another; the distribution of S cells in between is approximated with a set of harmonic functions. This theoretical histogram is subsequently disturbed with Gaussian dispersion functions to stimulate the pollution, yielding a predicted histogram. Using a maximum likelihood estimation technique, the model parameters are adjusted iteratively, matching the predicted histogram to the actually observed one. With the final parameter values substituted, the corresponding final theoretical histogram is regarded as a reliable reconstruction of the true histogram. From the latter, the required percentages can be read directly. The advantage of this approach over other mathematical analysis methods is that it allows a wide range of different, continuous distributions for relatively few model parameters (thus featuring flexibility and realism and a diminished risk of encountering computational problems). In addition, estimation errors providing a measure of accuracy can be obtained. To test the method, it was used to analyze various observed histograms from the literature that have been obtained by either simulation or actual flow cytometric measurements. The method appeared to perform well, as compared to the reported results of several other methods of analysis applied to the same data.     

Interactive multi-window integration of two-parameter flow cytometric data fields

Integration is necessary to determine the particle content of regions of interest of flow cytometric two-parameter fields. The improved program of the Cytomic 12 analyzer (1) offers: window trace integration for relatively simple window structures. The field of interest is surrounded by an integration trace (window). Eight independent windows can be stored and successively evaluated. It also offers painted field integration for complicated window structures. The pointer or a small window is interactively moved over the structures to be integrated like the brush of a painter. The "painted field" defines the window to be integrated. Window sets and painted fields can be stored on a floppy disk. Painted fields can be added and may also serve as look up tables for sorting.     

Use of a neural net computer system for analysis of flow cytometric data of phytoplankton populations

Flow cytometry has been used over the past 5 years to begin detailed exploration of the distribution and abundance of picoplankton in the oceans. Light scattering and fluorescence measurements on individual plankton cells in seawater samples allow construction of population signatures from size and pigment characteristics. The use of "list mode" data has made these studies possible, but on-shore analysis of copious data does not permit on-site reexamination of important or unexpected observations, and overall effort is greatly handicapped by data analysis time. Here we describe the application of neural net computer technology to the analysis of flow cytometry data. Although the data used in this study are from oceanographic research, the results are general and should be directly applicable to flow cytometry data of any sort. Neural net computers are ideally suited to perform the pattern recognition required for the quantitative analysis of flow cytometry data. Rather than being programmed to perform analysis, the neural net computer is "taught" how to analyze the cell populations by presenting examples of inputs and correct results. Once the system is "trained," similar data sets can be analyzed rapidly and objectively, minimizing the need for laborious user interaction. The neural network described here offers the advantages of 1) adaptability to changing conditions and 2) potential real-time analysis. High accuracy and processing speed near that required for real-time classification have been achieved in a software simulation of the neural network on a Macintosh SE personal computer.     

Sensitivity of flow cytometric data to variations in cell cycle parameters

Abstract We investigated to what extent flow cytometric DNA histograms are informative of cell cycle parameters. We created a computer program to simulate cell cycle progression in a generic and flexible way. Various scenarios, characterized by different models and distributions of cell cycle phase transit times, have been analysed in order to obtain the percentages of cells in the different cell cycle phases during exponential growth and their time course after mitotic block.
Cell percentages during exponential growth were insensitive to intercell variability in phase transit times and thus can be employed to estimate the relative mean phase transit times, even in the presence of non-cycling cells. However, this information is ambiguous if re-entry of such cells into the cycling status is permitted. The stathmokinetic outline gives the mean phase transit times, but also provides information about the spread, but not the form, of the phase transit time distributions, being particularly sensitive to the spread of G1 phase duration. The stathmokinetic outline also helps distinguish between scenarios considering only cycling cells, those forecasting a fraction of definitively non-cycling cells and those admitting a Go status with first-order output kinetics.     

Sensitivity of flow cytometric data to variations in cell cycle parameters

We investigated to what extent flow cytometric DNA histograms are informative of cell cycle parameters. We created a computer program to simulate cell cycle progression in a generic and flexible way. Various scenarios, characterized by different models and distributions of cell cycle phase transit times, have been analysed in order to obtain the percentages of cells in the different cell cycle phases during exponential growth and their time course after mitotic block. Cell percentages during exponential growth were insensitive to intercell variability in phase transit times and thus can be employed to estimate the relative mean phase transit times, even in the presence of non-cycling cells. However, this information is ambiguous if re-entry of such cells into the cycling status is permitted. The stathmokinetic outline gives the mean phase transit times, but also provides information about the spread, but not the form, of the phase transit time distributions, being particularly sensitive to the spread of G1 phase duration. The stathmokinetic outline also helps distinguish between scenarios considering only cycling cells, those forecasting a fraction of definitively non-cycling cells and those admitting a G0 status with first-order output kinetics.     

Improved method for computing potential doubling time from flow cytometric data

Relative movement methods use the timed progression of the mean fluorescence of cells which have been labeled with monoclonal antibodies against bromodeoxyuridine and displayed with bivariate flow cytometry according to DNA and label content to compute duration of DNA synthesis, TS. The relative movement is the difference of the mean DNA fluorescence of the labeled undivided cells from the G1 channel relative to the difference between the G1 and G2M channels. In this communication, we show how to extend this method to compute the potential doubling time, Tpot, the time required for a population of cells to double, given quiescent cells but no cell loss. A quantity v is introduced that is a function of the fraction of labeled divided cells and the fraction of labeled undivided cells. We show that v is independent of time and is equal to ln(2)Ts/Tpot so that Tpot (equal to ln(2)Ts/v) can be directly found from the information available in computing the relative movement. The method is applied to Chinese hamster ovary cells to demonstrate its utility.     

Identifiability of DNA distribution from flow cytometric data with cell debris

In flow cytometric measurement of cell DNA distribution one of the major problems is accounting for the effect of fragmentation in the staining process. This work considers a recent probabilistic model that has been proposed for the fragmentation process and species under which conditions it is possible to uniquely identify the DNA distributions of the original population using flow cytometric data. Attention is given both to the normal and to the polyploid case. This work was partially supported by a grant of the Italian National Research Council, Special Project “Oncology”, contract number 84.00632.44.     

Analysis of DNA synthesis rate of cultured cells from flow cytometric data

The rate of DNA synthesis along S phase is estimated from flow cytometric histograms on the basis of a mathematical model of a cell population. In the absence of loss, the model expresses the population kinetics in terms of DNA synthesis rate, S-phase influx, and population size. A single histogram is sufficient to determine the DNA synthesis rate when the population is in balanced exponential growth. Two suitably chosen histograms are necessary if the S-phase influx is exponential in a time interval longer than the S-phase duration. The analysis procedure was tested on published autoradiographic data and applied to three cultured cell lines (CM-S, 3LL, and M14 cells) that show various patterns of DNA distribution. In each case the cell-cycle fractions, the DNA synthesis rate, and the S-phase duration were obtained.     

Automatic identification of algae: neural network analysis of flow cytometric data

The performance of an artificial neural network for automaticidentification of phytoplankton was investigated with data fromalgal laboratory cultures, analysed on the Optical PlanktonAnalyser (OPA), a flow cytometer especially developed for theanalysis of phytoplankton. Data from monocultures of eight algalspecies were used to train a neural network. The performanceof the trained network was tested with OPA data from mixturesof laboratory cultures. The network could distinguish Cyanobacteriafrom other algae with 99% accuracy. The identification of specieswas performed with less accuracy, but was generally >90%.This indicates that a neural network under supervised learningcan be used for automatic identification of species in relativelycomplex mixtures. Incorporation of such a system may also increasethe operational size range of a flow cytometer. The combinationof the OPA and neural network data analysis offers the elementsto build an operational automatic algal identification system.     

Comparison of flow cytometric data obtained using fresh and paraffin-embedded lymphoid tissue

Flow cytometric (FCM) DNA analysis was carried out on 24 lymph nodes: 13 from benign reactive hyperplasias and 11 from non-Hodgkin's lymphomas. FCM was performed on two types of samples: (1) fresh cell suspensions and (2) suspensions prepared from formalin-fixed, paraffin-embedded sections. FCM of fresh samples detected aneuploidy in 23 of the 24 cases while FCM of paraffin-embedded samples detected aneuploidy in only 6 of the 24 cases. Those six cases were lymphomas considered histologically as having a poor prognosis. Only one case, a lymphoma, was euploid with both methods. The coefficients of variance determined in each case for both methods were found to be within "normal ranges," but were greater in the paraffin-embedded specimens. The results suggest that FCM DNA analysis of formalin-fixed, paraffin-embedded sections does not have as great a resolution capacity as does analysis of fresh cell suspensions, since the former failed to detect cell populations that had a small degree of aneuploidy (close to the 2n population).     

A new spreadsheet method for the analysis of bivariate flow cytometric data

Background  

A useful application of flow cytometry is the investigation of cell receptor-ligand interactions. However such analyses are often compromised due to problems interpreting changes in ligand binding where the receptor expression is not constant. Commonly, problems are encountered due to cell treatments resulting in altered receptor expression levels, or when cell lines expressing a transfected receptor with variable expression are being compared. To overcome this limitation we have developed a Microsoft Excel spreadsheet that aims to automatically and effectively simplify flow cytometric data and perform statistical tests in order to provide a clearer graphical representation of results.     

System for acquisition and real-time processing of multidimensional slit-scan flow cytometric data

The high-speed sampling requirements of multidimensional slit-scan signals (cell contours) have typically required custom hardware. This specialized hardware has often lacked the flexibility to adapt to varying instrument setups and experimental requirements. A hardware and software system capable of sampling multiple slit-scan cell contours at rates of up to 40 MHz with 10-bit resolution is described. It utilizes commercially available CAMAC transient recorders, a Digital Equipment Corp. PDP-11/83 computer, and custom hardware for signal conditioning and trigger generation. The modular design of the software system allows various hardware options with minimal additional coding. Real-time digital processing checks each cell contour for multiple peaks; extracts morphological features such as width, height, and area; accumulates gated histograms of these data; and optionally saves the derived data, selected contours, or both into list mode files on disk.     

A computer model of spermatogenesis in the rat; correlation with flow cytometric data based on autoradiographic cell-cycle properties

A computer model of rat spermatogenesis was created, based on autoradiographic studies of durations of the phases of the cell cycle (G1, S, G2 and mitotic phases) of each germ-cell type. With this model it is possible to predict and to gain insight into the changes of the DNA content occurring during the normal process of spermatogenesis. The relative proportions of haploid, diploid, S phase and tetraploid germ cells with increasing age of the rats were calculated. Calculated and actual experimental flow cytometry data were compared to test the accuracy of the model, and these show good agreement. The present work demonstrates that single-parameter DNA analysis of testicular cells is primarily a reflection of germ cells in the spermatocyte and spermatid stages of development, and of non-germ cells. The FCM single-parameter DNA analysis of testicular cells is relatively insensitive to changes in the stem cell and spermatogonial stages of germ-cell development.     

Realtime acquisition, storage, and display of correlated three-parameter flow cytometric data

A microprocessor-based system which performs realtime correlated acquisition, storage and display of multiparameter (3-parameter) data from a flow cytometer (FACS-III) is presented. List-mode techniques are not employed. The 3-parameter data is collected and correlated, then displayed along with cell-frequency as a realtime 3-parameter colour scattergram, while the experiment is in progress; in addition, correlated and uncorrelated higher-resolution projections of the 3-parameter data are collected and stored. The data projections may also be displayed: as 1-parameter histograms, or as 2-parameter colour or grey-scale scattergrams. Examples of 2- and 3-parameter colour scattergrams are presented. The speed and some characteristics of the realtime acquisition and display software are examined; methods to increase the realtime speed are discussed.     

Automated identification of subpopulations in flow cytometric list mode data using cluster analysis

The application of K-means (ISODATA) cluster analysis to flow cytometric data is described. The results of analyses of flow cytometric data for mixtures of fluorescent microspheres and samples of peripheral blood mononuclear cells are presented. A method for simultaneously displaying list mode data for any number of parameters, which had previously been applied to a continuous set of parameters such as multi-angle light scattering data, is used to present the results of cluster analysis on physically unrelated parameters; this method allows rapid evaluation of the success of subpopulation identification. The factors that influence automated identification of subpopulations are examined, and methods for determining optimal values for these factors are described.