Topic : Fluorescence compensation overview


STEP by STEP
       Flow cytometry relies in many aspects on the ability to detect fluorescent dyes, or "fluorochromes" which are bound in or on particles of interest. These fluorochromes have the universal characteristic that when excited by a light source, normally a laser, they emit energy as light at longer wavelengths. These emissions are rarely spectrally discrete and therefore overlap to some degree, so in order to measure the amount of signal in a given detector, a compensation network is employed to derive the appropriate amount of signal which should appear in each detector.

example of compensation from under to over..          Consider the animated picture to the left. It shows three populations of beads, one unlabelled and one each labelled with FITC and PE. Watch as the red and green coloured populations move from an uncompensated to a correctly compensated, then an over compensated position relative to the X and Y axes.

         So when is a population correctly compensated?
The short answer is when the median of a population positively labelled for Y is the same as the median of a population negatively labelled for Y when displayed relative to the X axis.

         Now let's use the simulator to adjust for spectral overlap on some real data files. The simulator will work with any FCS compliant data file, so you can practice and become comfortable with this concept using any file in the database, however if you'd like to use some data files which contain completely uncompensated data please select the following files:

Unstained:
FITC only:
PE only:
PerCP only:
APC only:

STEP by STEP

The basic procedure to follow is:

  1. select the file "beads_unstained";
  2. display plots of various fluorescence parameters;
  3. ensure that you have at least displayed plots which show fluorochromes which are spectrally close e.g. FITC & PE;
    Information on this and other combinations of fluorochromes can be viewed at TSRI Java spectrum viewer;
    Note that quadrant statistics are now available below each plot by selecting the appropriate item from the pop-up menu.
  4. next select and display a positively labelled sample, in this case we recommend the file "beads_FITC";
  5. select the "Adjust" menu and the Compensation item
    Selecting the compensation system
  6. Note that as soon as Compensation is selected, "slider bars" appear on the X and Y axes of the plot and "crosshairs" become active when the mouse is moved across the active plot.
    example of crosshairs on a plot example of locked crosshairs on a plot

    Position the crosshairs such that they suround the unstained population and click to "lock" the crosshairs (above right). Once the quadrant crosshairs are locked, statistics are generated for the active plot and they appear in the main statistics window

  7. The main statistics window shows arithmetic means and median value for each quadrant Upper Left (UL) etc. For these types of analysis, the median fluorescence intensity is a suitably robust statistic to give an a good indication of whether the compensation is correctly applied.
  8. Adjust the slider on the axis of the parameter you wish to compensate. For example, the slider which appears on the X axis of the dotplot when moved to the right will "subtract" an increasing amount of parameter X from parameter Y
  9. Remember the object of compensation is that parameter X -positive cells (e.g. FITC, FL1-H) should look no more positive in parameter Y (e.g. PE, FL2-H) than do negative cells.
A good web based discussion of fluorescence compensation can be found at the
Applied Cytometry Systems web site.
Other information can be obtained from Mario Roederer's web site.

The main reference for the logic behind N colour compensation is
"Fluorescence Spectral Overlap Compensation for Any Number of Flow Cytometry Parameters". C Bruce Bagwell and Earl G. Adams. 1993 Annals of the New York Academy of Sciences, Clinical Flow Cytometry, Volume 677

Also worth a look if you are interested in an "early" reference about fluorescence compensation is Loken, M.R et al. 1977 J.Histochem.Cytochem. 25:899-907