PRINCIPAL COMPONENT/BIPLOT ANALYSES FOR CELL SORTING

点击翻译按钮获取中文摘要

Multiparameter flow cytometry (FCM) data will be transformed during experiments to produce (using either principal components or projections based on other projection pursuit methods) two dimensional projections of the multidimensional data for visualization and sorting of cell subpopulations. Biplots, a powerful new technique for visualization of the correlation between multiple FCM parameters of each of the subpopulations will be used to see fundamental relationships between biological parameters which may give new insights as to basic biological mechanisms. The coefficients of the most useful projections of the data (based on linear or non-linear combinations of raw parameters) will be downloaded into real-time signal processing electronics. The raw parameters will then be transformed in real- time to the appropriate principal component (or other projection) coordinates. This will permit real-time scattergrams and cell sorting of cell subpopulations as seen by clusters of cell subpopulations in the projection space as well as on the basis of raw parameters which remain correlated at all times. The ability to physically sort out these cell subpopulations on the basis of principal component/biplot analyses can provide the critical link between subpopulations as defined by complex flow cytometric measurements and their biological importance as defined by functional and other assays that we will be performing on these isolated cells. These new methods will be specifically applied to complex multicolor fluorescence data from the blood of pregnant mothers. Additionally it will be wedded to the output of special pre- existing high-speed rare-events circuitry to allow improved multiparameter detection and sorting of rare human fetal cells from maternal blood. New methods for dealing with outlying data points should allow for improved detection of rare cell subpopulations normally obscured by "false positives". These new methods are of general utility and could provide other researchers using flow cytometry with powerful new methods for visualization of complex, difficult to interpret, data and a means of separating out new cell subpopulations of fundamental importance.