Quantitative Flow Cytometry: Instrument … Flow Cytometry: Instrument Characterization, Standardization and Calibration James Wood, PhD Manager, Wake Forest University School of Medicine

Quantitative Flow Cytometry: Instrument Characterization,

Standardization and CalibrationJames Wood, PhD

Manager, Wake Forest University School of MedicineComprehensive Cancer Center Flow Cytometry Shared Resource

Characterizing Cytometer Performance• Why characterize the cytometer

performance?• What can you characterize about a

cytometer’s performance?• Develop a consensus on which of the

characteristics are most relevant and how best to measure them?

Why Characterize Cytometer Performance?

• Better reproducibility of data• Facilitates intra/inter-lab comparisons of

cytometry data• May be required for clinical lab certification• Quantitative analysis of data• Data presented in physical units

What Can You Characterize about a Cytometer?• Precision• Linearity• Dynamic range• Sensitivity• Optical & Electronic Noise• Calibration in Spe Units

Precision• Alignment of laser and sample stream• Fluidics or pneumatics• Sample injection needle positioning• Bubbles and debris in the flow cell• Laser beam geometry• Signal attenuation• Photonic/electronic noise

LinearityCytometers must not only be linear but also

be proportional.Channel

Intensity Level 1 1x 100Intensity Level 2 2x 200etc. . . .

Dynamic Range

810121416182022242628

3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8

ADC

Reso

lutio

n

Decades

Required ADC Resolution (Bits)

Standardized Light Sources

• Fluorescent multilevel beads are traditional light sources

• Pulsed LED is a new light source with a low intrinsic CV

Beads vs. LED PulsesSpheroTech Bead Data

FITC-A

Cou

nt

-102-101 102 103 104 1050

40

79

119

158SpheroTech Bead Data

FITC-A

Cou

nt

-102-101 102 103 104 1050

40

79

119

158LED Data

FITC-A

Coun

t

-102-101 102 103 104 1050

1725

3450

5174

6899

Pulsed LED

Courtesy of Steve Perfetto, NIH/VRC

Prototype Pulsed LED

Sensitivity with Q & BThis spreadsheet calculates the detection efficiency, Q, and background, B, from the mean linear fluorescence and CV of a set of 3 beads. The beads, such as Rainbow beads, should be sufficiently dim in fluorescence that the CV's are broadened by photoelectron statistics and background light.Do not use data from a blank bead since the CV may be significantly in error.The CV of a bright, uniform bead corrects for intrinsic CV due to illumination and uniformity of staining.The mean linear fl units and MESF for a calibration bead are entered below: Enter data into yellow shaded areas. Calibration Bead Mean= 277 (linear FL units)

Calibration Bead MESF= 1400 (MESF)Bright Bead CV = 3.5 %

DATA for 3 beads CALCULATED VALUES Fit all three beads Fit only two dimmest beadsBead number Mean (linear fl

units)CV (as %)

F (MESF) Photoelectron SD Squared

1/Q B/Q 1/Q B/Q

1 146.9 10.75 742.5 5695.0 7.136860762 467.80197 Q= 0.1401 7.30563 270.9 Q= 0.13692 359.4 7.3 1816.5 13541.3 Std error 0.061350454 152.6873 B= 66 B= 373 759.4 5.58 3838.1 27821.9 R 2̂|seY 0.999926109 136.37551

13532.52634 1251681707.5 18598.28

Theoretical Equations relating CV, F, Q, B, and CVi CV^2F^2= B/Q + F/Q + F^2CVi^2CV^2= (B/Q)(1/F)^2 + ((1/Q)(1/F) + CVi^2

SDphotoelectron^2 = F^2(CV^2 - Cvi^2) = B/Q + F/Q

Plot SD-squared vs MESF

y = 7.1369x + 467.8020R2 = 0.9999

0.0

5000.0

10000.0

15000.0

20000.0

25000.0

30000.0

0.0 500.0 1000.0 1500.0 2000.0 2500.0 3000.0 3500.0 4000.0 4500.0

F (MESF)

SD S

quar

ed

Beads 1-3

Linear fit to data

Optical & Electronic NoiseBackgound Noise

FITC-A

Cou

nt

-102 -101 102 103 104 1050

95

191

286

381

M1

Parameter Arithmetic Mean

Standard Deviation

FITC-A 2.85 30.93

Calibration in Spe Units

y = 3E-05x2 + 98.669x + 4759.5

0.00E+00

5.00E+02

1.00E+03

1.50E+03

2.00E+03

2.50E+03

0.00

5000000.00

10000000.00

15000000.00

20000000.00

25000000.00

0 50000 100000 150000 200000 250000

Spe

SD^2

Intensity

26-Mar-2015 -

SD^2

Spe

Poly. (SD^2)

Linear (Spe)

Pulsed LED(light detection)

Calibration of known # of FL molecules per

mAb

Standard Beads(known # FL molecules)

Future of Detection

1

2

3

Courtesy of Steve Perfetto, NIH/VRC