Transcript No Slide Title

Digital cytometer setup and QC
Holden T. Maecker
Outline
• What do we care about in terms of cytometer
performance?
• What factors influence performance?
• What can we measure and track over time?
• What software adjustments are available?
• What QC parameters to look at:
• Initially on instrument setup
• Periodically (daily)
• With each experiment
What’s important to cytometer performance?
• Resolution Sensitivity:
• Q = optical detection efficiency
• B = background
• Linearity:
• Relative ability to detect fluorescence
differences across the detection scale
• Stability over time
Factors influencing sensitivity, linearity, and
stability
• Fluidics:
• System pressure and stability
• Optics:
• Laser type and performance
• Optical path, filters
• Electronics:
• PMT type and performance
• Signal processor performance
What can we measure and track?
• A set of beads of multiple intensities:
• Resolution of dimmest bead populations
and/or dim bead CV (sensitivity)
• Brightness ratio of two bead populations
across detection scale (linearity)
• Voltage required to achieve a certain bead
brightness (stability)
• Old: 8-peak Rainbow beads (Spherotech)
• Diva 6.0: CS&T beads (BD)
Standard bead sets
8-peak Rainbow beads:
CS&T beads:
Bright
2 µm
3 µm
Dim Mid
FSC
FITC
What adjustments are available?
• Laser delay
• Window extension
• Area scaling factor
What is laser delay?
Time from intersection
of 1st to 2nd laser
Time from intersection
of 1st to 3rd laser
• Dependent upon sheath velocity
• Can be mitigated by larger window extensions
What is a window extension?
Time window over which a pulse signal is measured
• Zero is the narrowest window, two is default
• Larger window extensions can overcome
minor imprecisions in laser delay setting
What is “area scaling factor”?
• A factor that is applied to the pulse area
measurement to equalize the height and area
scales
• Should rarely need to be adjusted
• Is important to keep bright events on-scale in
both area and height
What to measure and when?
• Initial instrument characterization
• Periodic (daily) performance check
• Experiment-specific setup
Initial instrument characterization
• Objective 1: characterize instrument sensitivity
and linearity
• Old method:
•
•
•
•
Run 8-peak rainbow beads
Set voltages so brightest bead is ~105
Observe separation in each detector (no std)
Manually adjust laser delay & area scaling factor
• Diva 6.0:
• Run baseline optimization routine
8-peak Rainbow beads
CS&T Baseline Optimization Report
Initial instrument characterization
• Objective 2: find baseline PMT voltages that
maximize resolution sensitivity
• Old method:
•
•
•
•
Run peak 2 beads at a series of voltages (350-800 volts)
Export CV of tightly-gated singlet beads for each detector
Plot CV versus voltage for all detectors
Lowest plateau voltage is the minimum baseline voltage for
that detector
• Run mid-range (peak 3) beads at minimum baseline
voltages, record mean in each detector (these are your
initial baseline target values)
• Diva 6.0:
• Part of baseline optimization routine
Manual determination of baseline PMT
voltages
Automated baseline PMT voltage
determination in Diva 6.0
Determining Baseline PMT Voltages
Baseline PMTV is set by placing the dim bead MFI to equal 10X SDEN
460 V
SDEN
Periodic (daily) performance check
• Manual method:
• Open Exp 1 (8-peak beads)
• Run a new tube of 8-peak beads with
same voltage settings
• Compare histograms to initial run
• Track at least one peak from each detector
for MFI and CV over time
• Diva 6.0:
• Run system performance check
Performance Tracking
• All measured performance parameters are
tracked:
• Linearity, CVs, Q and B, laser alignment
• PMT voltages required to hit target values
• Data is analyzed in Levey-Jennings plots
PMT Voltage
550
FITC Channel (Blue laser)
525
500
475
450
425
400
10/22/04 11/11/04 12/01/04 12/21/04 01/10/05 01/30/05 02/19/05 03/11/05
Time
Experiment-specific setup for a new panel
• Set voltages to achieve baseline target values
• Run single-stained CompBeads to see if each bead
is brightest in it’s primary detector
• If not, increase voltage in the primary detector (!)
• Run fully-stained cells and:
• Decrease voltages for any detectors where events
are off-scale
• Increase voltages for any detectors where low-end
resolution is poor (necessary?)
• Re-run single-stained CompBeads and calculate
compensation
• Record experiment-specific target values.
• Run samples.
CompBeads as single-color controls
CompBeads provide a convenient
way to create single-color
compensation controls:
• Using the same Abs as in the
experimental samples
• Creating a (usually) bright and
uniform positive fluorescent peak
• Without using additional cells
Experiment-specific setup for existing panel
• Set voltages to achieve experiment-specific target
channels.
• Run single-stained CompBeads and calculate
compensation.
• Run samples.
Why not put unstained cells in the first log
decade?
• Autofluorescence varies by detector
(very low in far-red range of spectrum)
• Leads to highly variable setup
• No guarantee that performance on
stained cells will be optimal
Why use target channels rather than voltages?
Once optimum voltages for a particular
experiment are determined, the settings
can be captured as target channels
(median of mid-range beads in each
detector).
These target channels are more robust
to instrument changes than are the
voltages themselves.
Summary
• We want our instruments to have good
sensitivity, linearity, and stability.
• Many factors contribute to sensitivity, linearity,
and stability, but we can track the net results
using a set of standard beads.
• Good setup and QC involves initial instrument
characterization, daily performance checks,
and experiment-specific setup.
• The CS&T module in Diva 6.0 automates
much of the QC workflow.
Diva 6.0: Making
instrument setup
and QC so easy, a
child can do it.*
*Not an official BD claim. The child
shown here has never run Diva 6.0,
but she’s watched her daddy do it.