DEVELOPMENT OF DATA DISPLAY &ANALYSIS SOFTWARE
University Of Calif-Los Alamos Nat Lab, Los Alamos NM
Investigators
Linked publications & trials
Abstract
Using the fourth generation computer language IDL (Interactive Data Language - Research Systems, Inc., Boulder, CO.) we have developed an elaborate program "IDLYK" for the display and analysis of flow data. This program has many functions and capabilities. Previously, one of its most important capabilities was the ability to convert our LACEL data files - collected on the numerous, unique instruments at Los Alamos - to the FCS (Flow Cytometry Standard) file format. This permits the use of other commercial software packages for special purposes such as cell-cycle analysis or complex immunophenotypic analysis. As we have purchased more commercial instruments, which are used for more routine experiments or moved our custom instruments to the DiDAC system, this capability has become less important. IDLYK has numerous, closely inter-related displays and analysis capabilities not found in the commercial software packages. For instance it is possible to go direc t ly from doing basic cell-cycle analysis to clustering based on a competitive-learning neural net algorithm. A composite panel of any of the available graph types can be assembled. These include image display, line-plot, shaded surfaces, and three parameter displays as either dot-plots or 3D clouds. The primary advantage of the inter-related displays is that a region of interest (a gate) drawn in one display format, e.g. a 2D graph, is mapped correctly to a 1D graph. Finally, there is an extensive feature set for kinetic data processing. This portion of the program extracts the mean of one parameter (the analytical parameter) versus another parameter, usually time. The program permits gating on scatter parameters to select a general cell subset and then means Vs time can be calculated on multiple regions of interest (to select specific cell subsets) and output the means to a text file which is exported to a spreadsheet or data base program IDL has also been used to develop an entirely new type of flow cytometry data processing, display and analysis software. This is a program "DNASizer" which accepts the count-rate history output of a photon-counting detector used in the unique systems, developed at Los Alamos, which can measure the size of fluorescently labeled DNA fragments in flow. The output of a photon-counting detector is usually a logic pulse for every detected photon of light. This stream of fast logic pulses is usually processed by programmable counter, which records the number of detected photons in small time intervals (usually 25 - 100 microsecond intervals). A data set might consist of several million of these interval counts recorded in several hundred scans from a Multi-Channel Scaler card's memory. The first task of DNASizer is to concatenate the multiple scan files into a single primary data record. Once the complete data set is assembled it must be processed to register the individual events (corr esponding to individual DNA fragments), integrate the total fluorescence of each burst of photons from a discrete fragment and record the maximum burst amplitude and the duration of the burst. Once these primary data are extracted from the raw data record primary histograms (area, height and width) are generated and displayed. Other graphs, such as a contour plot of burst area versus burst height can also be displayed. After various displays are generated and the processing is satisfactorily completed (and assuming multiple peaks are present in the burst area histogram) multiple Gaussians can be fit to the burst area histogram to objectively characterize the observed intensity of fluorescence from specific sizes of DNA fragments. Least squares fitting can then be done on the measured signal size versus the expected fragment size and a correlation coefficient determined. The results of analyzing a data set can be saved as GIF files, for the primary histograms, textual information re garding the file analyzed, and analytical result such as the curve fitting to import directly into Microsoft Access. This program is in use on several fragment sizing cytometers.
View original record on NIH RePORTER →