2013 Duke CFAR Flow Cytometry Workshop

2013 Duke CFAR Flow Cytometry Workshop

Data Analysis

Results from Pre-Workshop Analysis

Comp Profile





Elements of Data Analysis• Compensation – electronic adjustment for spectral overlap

– When to compensate• Acquisition – if gating on #CD3+, requires compensation• Off-line

– Spillover• Biexponential Transformation• Gates• Analysis Regions• Backgating – used to tweak gates and analysis regions so as

to optimize response (maximize positive and minimize negative responses)

• Training





• Training

B (3.4%)

F (10.5%)

E (13.4%)

D (10.2%)K (9.4%)

G (16.9%)

C (3.1%) A (6.8%)

H (10.2%)

I (12.7%)

J (4.8%)

• Here labs are listed in order of their total TNFa response. It is visually apparent that, while all labs had overcompensation, it is worst in labs with the lowest cytokine responses.

Inaccurate Automated Compensation:Requirement for Manual Adjustment

JO AnalysisAutocomp FlowJo

A700-PCPCy5.5 = 29.32

JO AnalysisModified comp FlowJo

A700-PCPCy5.5 = 6

HM analysisDiva (Lab J)

CD28

PCP

-Cy5

.5

CD3 A700

JO AnalysisModified comp FlowJo

A700-PCPCy5.5 = 6&

PCPCy5.5-PEA610 = 235

Note: Green laser excitation for both PerCPCy5.5 & PEA610

Compensation: Inspect and Manually Correct as Needed

PE-PEA610= 12.87

PE-PEA610= 11

Auto Manually adjusted

EQAPOL: example of

compensation affecting cytokine results

Comp Profile

Orig

inal

Mat

rix(A

uto-

com

p)

“Cor

rect

ed”

Mat

rix(A

uto-

com

pw

/ M

anua

l tw

eaki

ng)

Note 1: Compensation pairs discussed during the call are marked with pink arrows. Red arrows indicate other compensation pairs I felt could benefit from manually tweaking compensation values.Note 2: flowjo automatically flags manual edits using red text; all other differences are flowjo doing weird rounding/display stuff (ex. for PEA610-PE “590” is really “59.36;” the value has not been modified… this drives me NUTS!

Original vs Manually-tweaked FlowJo Compensation Values

15

Compensation Cannot Correct Spreading Error





• Training

No Biexponential Transformation:Off-scale Negative Affects Gate Placement

0 102

103

104

105

<B515-A>: IFNg FITC

0

102

103

104

105

<G71

0-A>

: CD4

CY5

5PE

20.6

0 10 2 10 3 10 4 10 5

<B515-A>: IFNg FITC

0

10 2

10 3

10 4

10 5

<G71

0-A>

: CD4

CY5

5PE

41

Original gate Revised gate

IFNFITC

CD

4 PE

-C

y5.5

FlowJo v8.3.3 (Rm 120 G5): BiExponential Transformation of Specimen 1 Tube 1 (Unstim) CD4+ Gate





• Training

CIC Gating Panel: Gating Recommendations

CIC Gating Panel: Gating Recommendations (examples of adequate analysis)

CIC Gating Panel: Gating Recommendations (examples of inadequate analysis)



– Spillover• Biexponential Transformation• Gates• Analysis Regions• Backgating – used to tweak gates and analysis regions so

as to optimize response (maximize positive and minimize negative responses)

• Training

EQAPOL: example of backgates showing CD3 dim+ excluded from gate

BeforeBackgate

AfterBackgate

IFNgBackgate

CD3 AmCyan

Excl

usio

n

0.38 5.74

CD4 Gated CD8 Gated

5.230.27

IFNg PE-Cy7

CD4

PerC

P-Cy

5.5

CD8

APC-

Cy7

BeforeBackgate

AfterBackgate

A

B

BACKGATING: purity & recovery

Duke University Medical Center





• Training

Intra-Operator Comparison: Original Analysis

N=5 FTE analyzing8 stims12 colors

Intra-Operator Comparison: Original Analysis

N=5 FTE analyzing8 stims12 colors

Intra-Operator Analysis:12 Color ICS NM Analysis - CD3+ Lymphocytes Gated

0 102

103

104

105

<B515-A>: IFNg FITC

0

102

103

104

105

<G71

0-A>

: CD4

CY5

5PE

20.6

0 10 2 10 3 10 4 10 5

<B515-A>: IFNg FITC

0

10 2

10 3

10 4

10 5

<G71

0-A>

: CD4

CY5

5PE

41

Original gate Revised gate

IFNFITC

CD

4 PE

-C

y5.5

Intra-Operator AnalysisBefore & After Correcting CD4- & CD8- Gates

original

final

Created in V6.4.2Opened & copied in V6.4.6-looks correct

Created in V6.4.6Opened & copied in V6.4.6-looks bad

Intra-Operator Analysis:Same data file created in different FlowJo versions but pasted

from the exact same FlowJo File (preferences identical)

Intra-Operator Analysis Before & After FlowJo Manual Transformation

Intra-Operator Comparison:Functional

Values

Gating Strategy

FSC-W

FSC-

H

88.3

<V705-A>: CD8 Q705

<G71

0-A>

: CD4

CY5

5PE

57.8

36.3

0.79

FSC-A

SSC-

A

99.3

<Violet G-A>: CD3 Amcyan

<Vio

let H

-A>:

vAm

ine

CD14

PB C

D19

PB

41.4

Gating Strategy for 11-Color Maturation/Function Panel: 1 of 3

CD4

PerC

P-Cy

5.5

SSC-

A

Excl

usio

n (V

iole

t H)

FSC-

H

FSC-ACD3 AmCyanFSC-W

CD8 Alexa700

Ungated Singlets CD3+ Exclusion-

Scatter

Basic Gates:

CD4+CD8-

CD8+CD4-

CD4+CD8+

- 3 total


<G66

0-A>

: CD2

7 CY

5PE

43 54.1

2.580.33

<G66

0-A>

: CD2

7 CY

5PE

56.4 28.6

8.466.55

<V54

5-A>

: CD

57 Q

545 0.12 1.07

55.942.9

<V54

5-A>

: CD5

7 Q

545

5.67 13.2

24.256.9

Gating Strategy for Sampson 11-Color Maturation/Function Panel: 2 of 3

<G66

0-A>

: CD2

7 CY

5PE

22 62.5

11.73.98

<V54

5-A>

: CD5

7 Q

545 3.98 22.9

51.721.5

CD57

FIT

C

CD57

FIT

C

CD57

FIT

C

CD27

APC

-Ale

xa75

0

CD27

APC

-Ale

xa75

0

CD27

APC

-Ale

xa75

0

CD45RO ECD

N

NN

CM

CMCM

EM

EM

EM

TE

TETE

E

EE

Maturational Gates:

CD4+CD8-

CD8+CD4-CD4+CD8+

CD45RO ECD

CD45RO ECD

Naive Central Memory

EffectorMemory

Terminal Effector Effector


EffectorMemory



EffectorMemory


- 5 per basic subset


<R710-A>: CD107a AX680

2.59CD107


Functional & Boolean Gates:

- 4 functional gates per maturational subset

CM: CD8+CD4-

1.14

IL-2

TNF-a

IFN-

0.31

4.19


Backgate!

<R710-A>: CD107a AX680

2.59CD107


Functional & Boolean Gates:

- 4 functional gates per maturational subset - 16 boolean gates per maturational subset

CM: CD8+CD4-

Boolean Gates

Polyfunctional (1: ++++)

Polyfunctional (4: +++)

Bifunctional (6: ++)

Monofunctional (4: +)

Nonfunctional (1: ----)

Key:7 = CD107g = IFN-2 = IL-2T = TNF-a

1.14

IL-2

TNF-a

IFN-

0.31

4.19


2013 Duke CFAR Flow Cytometry Workshop

Documents

used singlet gate

iifng vs cd3 gate

amine gate

used scatter gates

vs amine

time gate time gates

cd3 dim

cd4 vs cd8