Numerous statistics are produced for each sample within a batch. Definitions for these terms can found in the xPONENT® 4.0 software manual, and have been summarized here for your use!
When a sample is acquired, the fluorescence intensity (FI) observed for each bead is recorded. Median, Mean, Standard Deviation, Peak, as well as the Trimmed values (i.e. Mean, Standard Deviation, and Peak), are derived directly from the FI. Because %CV and Trimmed%CV are derived from Mean and Standard Deviation, they are also a function of FI.
Many of the statistics available in xPONENT depend on the settings of your protocol:
- Net MFI (Median Fluorescent Intensity) is the MFI minus the background if a background well has been selected within the protocol. If a background well is not selected in the protocol, then Net MFI = MFI.
- Normalized Net Median is dependent on whether a normalization bead is selected in the protocol. If it is, then the MFI for each bead set is divided by the MFI of the normalization bead set, giving Normalized Net Median.
- Control Range Low/High, Expected Result, Units and Range are defined by the protocol/standards/controls.
- %Recovery is a function of the analysis of standards. After data is collected for standards, the analysis software draws a best fit curve. Because we don’t live in a perfect world, no standard or control is expected to fall exactly on the best fit curve. The calculated (observed) value of your standard/control divided by the expected value x100% is the %Recovery.
- Count is the number of beads counted per analyte, and Trimmed Count is 90% of Count. Clumped beads are thrown out and do not contribute to count or FI, due to size exclusion limits defined in the doublet discriminator settings. Because beads from different sets clump together, they are scattered all over the bead map.
Want to see more than just the MFI?
Luminex long ago determined that the MFI calculated for each sample and analyte is best for analysis and reproducibility. Since the MFI is a median value, some statistical savvy folks wish to analyze the raw FI data. One can obtain the FI for each bead within each sample from a so-called rCSV file. To configure your instrument to automatically generate the rCSV files, simply go to the Admin page/CSV Options Tab and make sure the bottom-most checkbox is checked. If checked and a run is performed, the rCSV files will be located under:
C:/docs&settings/all users/application data/Luminex/xponent##/batchXXX_rcsv
(batchXXX is the name of your batch). The rCSV file can be viewed in Microsoft Excel.
Statistic Definitions (from xPONENT 4.0 Manual)
- Median – The point in a data set where there are as many values occurring above it as there are below it.
- Test Result – The calculated analysis value for quantitative or qualitative assays derived from standards with known values.
- Range – A sem-quantitative result for a particular numerical result falling between a predefined set of values such as “Normal” or “Negative”.
- Count – Gated events if gates are set.
- Net MFI (Median Background) – Net MFI median background.
- Mean – Average of all values in a set of results.
- % CV of Microspheres – The measure of relative dispersion within the distribution. %CV = 100 x Std Dev / Mean
- Standard Deviation – For calculating sample variability or dispersion, Luminex uses the standard deviation formula.
- Peak – The value that is equal to the largest number of data points within the distribution. For example, in data set {1,2,2,3,3,3,4,5}, 3 is the peak because it occurs the largest number of times in the distribution list.
- Trimmed Count, Trimmed Mean, Trimmed%CV of Microspheres, Trimmed Standard Deviation, Trimmed Peak – Trimmed statistics remove the lower and upper five percent of the extreme statistic values, then use the remaining values for the Mean, Standard Deviation, or %CV calculations.
- % CV of Replicates – The measure of relative dispersion within the distribution of results for replicate samples. %CV = 100 X Std Dev / Mean
- % Recovery – A measure of how accurately your observed results match your expected results following regression analysis. (Observed concentration) / (Expected concentration) x 100%
- Expected Result – The known or expected test result value for a standard or control.
- Control Range – Low – The lowest value for an assay control used to determine pass/fail criteria for an assay.
- Control Range – High – The highest value for an assay control used to determine pass/fail criteria for an assay.
- Normalized Net Median – For each analyte in a well the Normalized Net Median (NNM) = (net median of analyte) / (net median of normalization bead)
- Units – The unit of measure for an analyte, for example, pg/mL.