PROBABILITY STATE MODELING: A NEW PARADIGM FOR CYTOMETRIC ANALYSIS - Flow Cytometry in Drug Discovery and Development - page 281

FIGURE 14.7 How probability state modeling accounts for population overlap. Because
measurement error is represented on the y-axis (vertical arrows) and percentages are estimated
from the x-axis, probability state models account for population overlap due to measurement
error.

14.5.1 Two or More Progressions or Cell Types

What happenswhenwewish toevaluate twoormoreprogressions ina sample?Howcan
the probability state model include this additional level of classification in its algo-
rithms? Again, the answer is easier than you might think. When there is more than one
progression, the modeling system evaluates each progression or cell type separately. It
finds all the state probability weights for each cell type and then summarizes these
weights as a single cumulative weight. Another stochastic selection process uses these
cumulative weights to choose a cell type. Once the cell type or progression is chosen, it
then uses the same method as described earlier to find the state within that cell type.

14.5.2 It Is Too Good to be True

Wewill investigate this powerful capabilitymore in a fewmoments. Beforewe look at
how this technique works for some complex applications, it is important to clear up
one last issue. Those of you who have been reading along should have a nagging