PROBABILITY STATE MODELING: A NEW PARADIGM FOR CYTOMETRIC ANALYSIS - Flow Cytometry in Drug Discovery and Development - page 284

FIGURE 14.9 Sequence of analysis steps for finding the normal bone marrow B-cell lineage
in a 12-parameter listmode file. This figure shows the first four analysis steps (data kindly
provided by Fred Preffer, Harvard University). See text for details.

at SSC. We see some CD19 positive events in the granulocytes, so we include a SSC
constant parameter profile in the lymphocyte region to eliminate these events from
our B-cell selection (Figure 14.9b). Keep in mind that we could have started with
SSC and then done CD19 and the final solution would have been the same. Notice
that our frequency distribution (Figure 14.9, bottom panels) remains uniform as we
go through this process.
A parameter overlay plot can be used to show our progress thus far. CD19 and SSC
are essentially constant. The plots are bumpy because B-cells comprise only about
2.3% of the complete sample (Figure 14.9a and b bottom overlay plots) and counting
statistics is adding a lot of variability to our model at this point.
We can eliminate much of this statistical variability by enriching our sample for
just B cells. Notice that after enrichment our % classification becomes almost 100%
and the SSC and CD19 plots become much smoother (Figure 14.9c).

14.6.1.3 TdT Step-Down Parameter When we examine TdT, we initially see two
distributions in the data (Figure 14.9c). Note, only two parameter profiles are shown at
a time. The initial CD19 distribution has been collapsed in order to show the new
parameters at full scale. There are relatively fewTdT þ events and a lot of TdT events