center for intelligent chemical instrumentation 1 chemometric considerations in proteomic analyses...

Center for Intelligent Chemical Instrumentation 1

Chemometric Considerations in Proteomic Analyses by Mass

Spectrometry

Peter de B. Harrington* Mariela L. Ochoa*, Sanford P. Markey+, Claudine Laurent+, Kuniaki Saito+, & Alfred L. Yergey^

*Ohio University, Center for Intelligent Chemical, Instrumentation, Department of Chemistry and Biochemistry, Athens, OH 45701-2979, [email protected]

+Laboratory of Neurotoxicology, National Institute of Mental Health, Building 10 Room 3D42, MSC 1262, 10 Center Drive, Bethesda, MD 20892-1262

^Section on Mass Spectrometry and Metabolism, Building 10, Room 9D52, National Institute of Child Health and Human Development, 10 Center Drive, Bethesda, MD 20892-1580


Chemometrics

Chemometrics is a discipline that is devoted to maximizing the amount and quality of information obtained from chemical or molecular measurements.

Chemometrics uses mathematical, statistical, logical, and computational tools.

For scientists, an important chemometric topic is the statistical design of experiments.


‘Omics Era

• Besides the proteome over 50 other ‘omes*.

• Complex biological systems• Reductionism is difficult because of the

large degree of interaction.• The interesting proteins are the ones

that are difficult to detect.

*http://www.genomicglossaries.com/content/omes.asp, accessed on 16-Mar-2004.


Chemometrics and Proteomics

KnowledgeBasesBiology

Sample

Preparation

Instrumental Measurement

Data Information


Some Statistics Concerning Foodborne Bacteria Pathogens

• In the U.S., 76,000,000 foodborne illnesses occur each year (325,000 hospitalizations and up to 5,000 deaths).

• Escherichia coli O157:H7 foodborne poisoning:– Largest outbreak (1993): more than 700 people ill and 4 deaths– Up to 75,000 infections estimated annually

• Listeria monocytogenes foodborne poisoning:– Largest outbreak reported in 1985– About 2,500 cases of Listeriosis every year – 500 deaths attributed to Listeriosis

Buzby, J. C., Frenzen, P. D. and Rasco, B., Product Liability and Microbial Foodborne Illness. Agricultural Economic Report Nº 799; Food and Rural Economics Division, Economic Research Service, U.S. Department of Agriculture: Washington, DC. April 2001 p 1., Website http://www.about-listeria.com/aer799.pdf, (accessed Feb 2004).

http://www.about-listeria.com/aer799.pdf


IMS and MALDI TOF-MS as Attractive Methods for Foodborne Bacteria Characterization

Ion Mobility Spectrometry (IMS)

1. Presumptive technique

2. Ion mobility spectra may furnish useful information for bacteria species/ strains characterization and differentiation

3. Fast analysis time for rapid screening of foodborne pathogens

4. Portable instruments, attractive for on-site monitoring

Matrix-Assisted Laser Desorption/Ionization

(MALDI) TOF-MS1. Confirmatory technique2. Provides a fingerprint of proteins

for bacteria of interest3. Comparison against database

containing the bacteria genome alleviates the issue with spectral reproducibility

4. Rapid analysis time


Identification of Foodborne Pathogens Using Molecular Weight Database Search

The ExPASy (Expert Protein Analysis System) proteomics server of the Swiss Institute of Bioinformatic (SIB) Home Page http://us.expasy.org/srs/ (accessed Oct 2003).

Database Search for Organism’s Protein Molecular Weight

http://us.expasy.org/srs/


Problems Associated with Microbiological Food Analysis

• Detection of small number of pathogens hampered by large numbers of harmless background microflora

• Culture enrichment steps necessary to amplify target analytes before traditional methods of detection can be applied

• Affinity capture techniques (i.e., immunomagnetic separations –IMS) to isolate target bacteria from complex food matrices

Madonna, A. J.; Basile, F.; Furlong, E.; Voorhees, K. J. Rapid Commun. Mass Spectrom. 2001, 15, 1068-1074.


MALDI as an Ionization Method

• Introduced by Karas and Hillenkamp (1987) as ionization method for non-volatile polar biological and organic macromolecules and polymers

• Low concentration of analyte uniformly dispersed in solid or liquid matrix

• Matrix should have strong absorbance at laser excitation wavelength and low sublimation temperature

• Three main processes occur: formation of solid solution, matrix excitation, and analyte ionization

Karas, M.; Bachmann, D.; Bahr, U.; Hillenkamp, F. Int. J. Mass Spectrom. Ion Process. 1987, 78, 53-68.


M@ldi-LRTM Mass Spectrometer Time-of-Flight by Micromass (UK)

Instrumental parameters

• Laser: Nitrogen UV (337 nm)

– Firing rate: 5 Hz

– 10 shots/spectrum

• Ion optics: Linear TOF path

length 0.7 m

• Ion source: Grounded “time lag

focusing” source (delayed

extraction) ~ 500 ns

• Accelerating voltage: 15 kV

• Detector: Fast dual micro-

channel plate (MCP)


MALDI-Time-of-Flight Mass Spectrometry (TOF-MS)


Variation in the MALDI Mass Spectrum

• Compare the signal averaged spectrum to a collection of single laser shots.

• Single scan spectra are from individual laser shots.

• Historically spectra were signal averaged because of computational limits on storing large amounts of data.

• Modeling the single scan spectra can be beneficial.


1 1.5 2 2.5

x 104

0

5

10

15

20

25

30

35

m/z

Inte

nsi

ty (

cou

nts

)

Average MALDI-MS Spectrum for a Protein Standard Mixture

Cytochrome cMyoglobinTrypsinogen



Baseline Correction

• Polynomial or exponential fitThe model usually depends on the instrument or matrix conditions

• Reduce the least squares error between the spectrum and the model

20 1ˆ b xy b b e


1 1.5 2 2.5

x 104

-5

0

5

10

15

20

25

30

m/z

Inte

nsi

ty (

cou

nts

)

Baseline Corrected MALDI-MS Spectrum for a Protein Standard Mixture




Wavelet Compression

• For large data sets, modest linear wavelet compression can improve efficiency.

• The biorthogonal wavelets, such as the Villasenor preserve the peak locations and avoid the extra step of reconstruction.

• Compressed using 4 levels and a biorthogonal filter with 3 vanishing moments.

• Improves signal-to-noise ratio by removing high frequency components.


0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6

x 104

-1

0

1

2

3

4

5

6

7

8x 10

4

m/z

Inte

nsi

ty (

cou

nts

)

Wavelet Compressed Average Spectrum

Average Spectrum 100KWavelet Coefficients 6K

1.4 1.42 1.44 1.46 1.48 1.5 1.52

x 104

-2000

0

2000

4000

6000

8000

10000

m/z

Inte

nsi

ty (

cou

nts

)

Wavelet Compressed Average Spectrum

Average Spectrum 100KWavelet Coefficients 6K


Modern Approach

• Compress single shot scans • Baseline correct• Align m/z drift for each individual

scan (i.e., single laser shot spectrum).

• Model using multivariate curve resolution


Multivariate Curve Resolution

• Simple linear models based on transient behavior of the data

• Separate correlated spectral information based on temporal response

• Simple-to-use interactive mixture analysis (SIMPLISMA).

• Alternating least squares (ALS)


Con

cen

trat

ion

Spectra

TD = CS E

Model = Product of analyte concentration and analyte

sensitivity

= +

Error


Principal Component Analysis• Decomposition into

orthogonal matrices C and S

• The matrices maximize variance

• The matrices are abstract in that they do not represent physical or chemical trends


SIMPLISMA

Willem Windig and Jean Guilment, Anal. Chem. 1991, 63, 1425-1432.

11 1,2

1

iii

i iii

r rp

r r

Instead of detecting peaks, SIMPLISMA selects points or columns in the data matrix D that have a maximum purity.

The two criteria for a pure variable are:

1. the point characterizes a variance

2. the point varies independently with other points in the model


SIMPLISMA Decomposition

• The columns of the data matrix D are used as initial estimates for the concentration profiles C.

• Spectra are obtained by least squares regression of C onto D.

• The spectra are normalized to unit vector length.

• Concentration profiles are obtained from regression of the normalized spectra S onto D.

2

T T -1

T -1

S = D C(C C)

S = S/ S

C = DS(S S)


Alternating Least Squares (ALS)

• Alternating procedure of regression with constraints

• Concentrations and spectra should not be negative. Use non-negative constrained least squares for the regression.

-1

-1

f ( )

f ( )

Tn+1 n

n+1 n+1

S = D C

C = D SJ. C. Hamilton and P. J. Gemperline, "Mixture Analysis Using Factor Analysis II: Self Modeling Curve Resolution," J. Chemometrics, 1990, 4, 1-13.


1 1.5 2 2.5

x 104

-0.1

-0.05

0

0.05

0.1

0.15

0.2

0.25

Mass- to-charge Ratio (m/z)

No

rmal

ized

In

ten

sity

Simplisma Spectra of Unaligned Scans

Component #1Component #2


2.38 2.4 2.42 2.44 2.46 2.48 2.5

x 104

0

0.05

0.1

0.15

0.2


No

rmal

ized

In

ten

sity

Simplisma Spectra of Unaligned Scans



2.39 2.4 2.41 2.42 2.43 2.44 2.45 2.46

x 104

0

20

40

60

80

100

120

140

160


Inte

nsi

ty (

Co

un

ts)

First 10 Unaligned Scans


Mass Alignment

• Before alignment each scan is wavelet compressed and baseline corrected.

• Align each spectrum so that the correlation with the average spectrum is maximized.

• The alignment is obtained by a quadratic fit of the m/z of each spectral scan.

• Linear interpolation is used to match the scan m/z to the average m/z of the data set.


1 1.5 2 2.5

x 104

-0.05

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4


No

rmal

ized

In

ten

sity

Simplisma Spectra of Aligned Scans



2.35 2.4 2.45 2.5

x 104

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35


No

rmal

ized

In

ten

sity

Simplisma Spectra of Aligned Scans



2.38 2.39 2.4 2.41 2.42 2.43 2.44 2.45

x 104

0

20

40

60

80

100

120

140

160


Inte

nsi

ty (

Co

un

ts)

First 10 Aligned Scans


0 100 200 300 400 500 600-500

0

500

1000

1500

2000

2500

3000

Scan Number

Inte

nsi

ty (

Co

un

ts)

Concentration Profiles of Aligned Scans



2.38 2.4 2.42 2.44 2.46 2.48

x 104

0

0.02

0.04

0.06

0.08

0.1

0.12


No

rmal

ized

In

ten

sity

Comparison of Signal Averaged and SIMPLISMA Spectrum

SimplismaProc-Mean


-1000 0 1000 2000 3000

-500

-400

-300

-200

-100

0

100

200

300

400

500

1

2345678 9101112

131415

16

171819

2021222324

2526

27

28

29

30

31323334

353637

383940414243444546474849

50

5152535455565758

596061626364656667686970717273747576 77

787980

818283848586878889

90 9192

9394959697

9899100101102103104105106107108109110

111112113114115116117118

119120

121

122123124125126127128129

130131132133134135136 137138139

140141142143

144145146147148149150151152153154155156157158159160161162163

164165166167168169170171172173174175176177178179180181182183184185186187188189190191

192193194195196

197

198199

200201202203204205206207

208209

210211212213214215216217218219220221222223224225226227228229230231232233234

235

236237238239240241242243

244245246247248249250251252253254255256

257258259

260261262263264265266267

268269270271272273

274275

276277

278279 280281

282

283284285286

287288

289

290291292293294295296297

298299

300301

302

303

304305306307308

309310

311

312313314315316317

318319320

321322323324

325326327328329

330331

332333334335336337338339340341342343344345346347348349350351352353

354355

356357358359360361362363364

365366367368369370371

372373

374

375376377378

379

380381382383384385

386387388

389390391392

393394395396397

398399400

401

402403

404

405

406

407

408409410411

412413414415416

417418419

420

421422423424

425426

427428

429

430431432

433

434

435

436

437

438

439

440441

442

443444

445446

447448449450451452453454

455456457458459460461

462463464

465466

467468469470

471

472473

474475

476477478

479480 481

482

483

484485

486487

488489

490491

492

493494

495 496

497498499

500501

502503504505506507508509510511512513514515

516517

518519520521522

523524525526

527528529

530531532533534535536537538539540541

542543

544

545

546

547548

549

550551

552

553

554

555

556

557558559

560561562563564565566

567

568569

570

571

572573

574575576577

578579

580581582

583584

585586

587

588589590

591

592593

594

595596

597

598

599600601

602

603604605606607608609610611

612

613614615

616617

618

619

620621622623624625626627628

A

PC #1 (81%)

PC #

2 (

2%

)PCA Score Plot for Processed MS

Plotted with respect to scan number.


-0.8 -0.6 -0.4 -0.2 0 0.2 0.4

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

2 3

456

78

91011

121314

1516

1718192021222324

252627282930313233343536

37383940

41

42

43

44

45

46

47

48 49 50

51

52

53

54

55

56 57585960

61

6263

64

65

66

67

68

69

70

71

72

73

74

75

76

77

7879

80

81

8283848586

878889

90

919293

94

95

96

97

9899100101

102103104105

106

107

108

109110111

112113

114115

116117118

119

120121

122

123

124 125126127

128

129130131132133134135136

137138139140141142143

144

145

146147148149

150

151152

153

154

155

156

157

158159

160161162

163

164

165

166

167

168

169170

171

172

173

174175176177

178179

180

181182183

184

185

186187

188

189

190

191 192

193194195

196

197

198

199

200201

202

203

204

205206207

208

209

210211212

213

214

215

216217

218

219

220

221

222

223

224225

226

227228

229

230231

232233

234 235

236

237

238

239

240

241

242243

244

245246

247

248

249

250251

252253

254

255

256257

258

259

260

261

262 263264

265

266

267

268

269

270271272273274

275276

277278279280281282283

284285

286287288 289290

291292293294295296297298

299300301302

303

304305

306

307308 309310

311312313

314

315316317

318319 320

321322323

324

325326

327

328

329 330

331

332

333334

335

336

337

338339

340

341342

343

344

345

346

347

348

349

350

351

352353 354355

356

357

358359

360361

362

363

364

365

366367368

369370

371

372373 374375376

377378379380381382383384385386387

388389390391392393394395

396397398399

400

401

402403

404405406407

408 409410

411

412

413

414

415

416

417418

419

420421422423424425426427428429430431432433

434435436437438439440441442443444445446447448449450451452453

454

455

456

457 458459

460

461

462463

464465

466467468

469470471472473474475476477478479480481482483484485486487488489490491492

493494495496497498

499500501502

503504505506507508

509510511512513514515

516517

518519520

521

522

523524

525526

527

528529

530

531

532 533

534

535536

537

538

539 540

541542543544545546547

548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578

579580581582583584585586587

588589590591592593594

595596

597598599600601602603604

605

606

607

608

609

610

611612

613

614

615 616

617

618

619

620621

622623

624

625626

627

628

A

PC #1 (17%)

PC #

2 (

7%

)

PCA Score Plot of Normalized Processed Scan


1 1.5 2 2.5

x 104

-0.05

0

0.05

0.1

0.15

0.2

0.25

0.3


No

rmal

ized

In

ten

sity

Variable Loadings of the First 2 Principal Components



1 1.5 2 2.5

x 104

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

0.18


No

rmal

ized

In

ten

sity

Comparison Between ALS and Mean Spectra

ALS SpectrumMean Spectrum


2.35 2.4 2.45 2.5

x 104

0

0.05

0.1

0.15

0.2


No

rmal

ized

In

ten

sity




1.4 1.42 1.44 1.46 1.48 1.5 1.52

x 104

0

0.005

0.01

0.015

0.02


No

rmal

ized

In

ten

sity




Prediction of Risk for Premature Delivery from MALDI-MS of Amniotic Fluid

• Control was a pooled amniotic fluid from women who produced excessive volumes of amniotic fluid (AF)

• Women who are at risk for premature delivery from two individuals

• Three replicates of each sample were studied at different times

• Each replicate was subject to one of four sample preparation procedures

• After sample preparation 3 more replicates were obtained to characterize measurement variations


Sample Preparation• The matrix was formed from saturated sinnapinic

acid in a 1:1 mix of acetonitrile (ACN) and 0.1% trifluoroacetic acid (TFA)

• Four sample preparation procedures were evaluated

• The samples were diluted 10-fold to volume with 0.1% TFA

• Method 1 adds a 1.0 L of this solution to the MALDI plate

• Method 2 extracts 15 L with a ZipTip and elute with 5 L of a 1:1 mix of TFA 0.1% and ACN

• Method 3 extracts a diluted 10-fold solution with 5 L of methylene chloride

• Method 4 is method 3 followed by method 2


MALDI-MS Conditions for Amniotic Fluid Study

• ABI Voyager DE-STR – Linear mode– Delayed extraction 375 ns– Accelerating voltage 25 kV– Grid 95%– Guide wire 0.1%– Mass range 3-20 kDa– Low mass gate 2 kDa– Laser shots per spectrum 250


Analysis of Variance (ANOVA)

• The data set comprised 108 spectra with 38,970 mass measurements

• Additive variance model coupled with PCA

Pr Pr

( ) ( ) ( )

( ) ( ) ( ) ( )Treatment Patient Treatment Sample Patient

ep Sample MS ep Int MS I nt

x x x x x x x x

x x x x x x x x


-0.6 -0.4 -0.2 0 0.2 0.4

-0.4

-0.2

0

0.2

0.4

0.6

1

2

34

56

78

9101112

131415

16

1718

19

2021

22

2324

252627

28

2930

313233

34 3536

3738

39

404142

434445 46

47

48

495051

525354

55 5657

5859

60

6162

63

6465

66

6768 69

7071 72

73 7475

76 7778

798081

82

8384

8586

87

8889

90

9192 93

94

9596

979899

100101

102

103104105

106107108

ControlPre

PC #1 (28%)

PC

#2

(24%

)Total Variance


-0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

1

2

34

56

78

9

10

1112

131415

16

17

18

19

202122

23242526

27

28

2930313233

34

35

36

37

38

39

40

4142

4344

45

46

47

48

49505152535455 5657

58

59

60

61

62

63

64

65

66

676869

70

7172

73

7475

76

7778

798081

82

8384

85 8687

88

89

90

91

92 93

94

95

96

9798 99

100

101

102

103104105

106

107

108

ControlPre

PC #1 (47%)

PC

#2

( 2%

)

Treatment vs Residual Variance


0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2

x 104

-0.1

-0.05

0

0.05

0.1

0.15

m/z

Rel

ativ

e In

tens

ityVariable Loadings of Treatment


-0.2 -0.1 0 0.1 0.2 0.3

-0.6

-0.5

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

12

3 45

6 789

10

1112

131415 1617 18 19202122 2324252627 282930 31323334

35 3637 3839 404142 4344

45

46

4748 4950

51 525354 555657 58

59

6061

6263 6465

66

676869 70

7172 737475 767778 798081

82

8384

85868788

89

90919293

9495

96 979899100

101102 103104105

106107108

ControlStudy 1Study 2

PC #1 (15%)

PC

#2

( 4%

)

Study vs Residual Variance


0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2

x 104

-0.04

-0.02

0

0.02

0.04

0.06

0.08

m/z

Rel

ativ

e In

tens

ity

Variable Loadings of Study


-0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4

-0.25

-0.2

-0.15

-0.1

-0.05

0

0.05

0.1

0.15

1

23

45

6

78

9101112

13

1415

16

1718

19

2021

22

2324

252627

28

2930

313233

343536

37

38

39

404142

434445

46

47 48

495051

52

5354

55

5657

58

5960

61

62

63

6465 66

676869

70

71

72

7374

75

76

77

78

79

8081

82

8384

8586

878889

90

91

92

9394

9596

97

98

99

100

101

102

103104105

106

107

108

Ex-ZTExZTNoZT

PC #1 (35%)

PC

#2

( 3%

)

Pretreatment vs Residual Variance


0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2

x 104

-0.04

-0.02

0

0.02

0.04

0.06

0.08

m/z

Rel

ativ

e In

tens

ity

Variable Loadings of Pretreatment


-0.6 -0.4 -0.2 0 0.2 0.4 0.6

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

12 3

456

789

10

1112

131415

161718

192021

222324

252627

282930

313233

34

3536

37

38

39

404142

4344

45

46

47 48

495051

525354

55

565758

59

60

61

62

63

64

65

666768

6970

71

72

73

7475

76

7778

79

8081

828384

8586

87

8889

90

91

92

93

94

9596

97

98

99

100

101

102

103104

105106

107108

Ex-ZTExZTNoZT

PC #1 (18%)

PC

#2

( 7%

)

Interaction vs Residual Variance


Follow-up ExperimentDay 1 Day 2 Day 3 SB1b-n PB2c-n SB3b-nPA1a-ZT SB2c-ZT PA3b-nPB1a-ZT SB2a-ZT PC3b-ZTPA1b-n SA2a-n SC3a-nPC1c-ZT PC2c-ZT SB3b-ZTSC1c-n PB2b-n SA3c-nPA1c-ZT PA2b-n SB3a-nSC1c-ZT PB2c-ZT SA3a-nSC1a-n SA2b-ZT SC3b-nSC1a-ZT PC2b-n PC3a-ZTPA1c-n SB2c-n SA3a-ZTSA1b-n PB2b-ZT SC3a-ZTSC1b-n PC2c-n SA3b-nSB1b-ZT SC2b-ZT PB3a-nPC1a-n SA2a-ZT SA3c-ZTPB1b-n SA2c-ZT SC3b-ZTPC1c-n SB2b-ZT PB3c-nSB1a-ZT PA2c-n PA3a-n

Day 1 Day 2 Day 3 PB1b-ZT SC2c-n PB3b-ZTSB1a-n SC2c-ZT PC3a-nSB1c-n PB2a-n PA3c-ZTSA1a-ZT SC2a-n PB3b-nSA1a-n PC2a-n PA3b-ZTPB1a-n PB2a-ZT PB3a-ZTSA1c-ZT PA2a-n SA3b-ZTPC1a-ZT SA2b-n SB3c-ZTPC1b-n PC2a-ZT SB3c-nSB1c-ZT SB2b-n PC3c-nPA1b-ZT SC2a-ZT PB3c-ZTPB1c-ZT PA2c-ZT SB3a-ZTPB1c-n SB2a-n PA3c-nPC1b-ZT SC2b-n PC3b-nSA1b-ZT PC2b-ZT PC3c-ZTSA1c-n PA2bZT SC3c-ZTSC1b-ZT PA2a-ZT SC3c-nPA1a-n SA2c-n PA3a-ZT


-0.1 -0.05 0 0.05 0.1 0.15 0.2

-0.25

-0.2

-0.15

-0.1

-0.05

0

0.05

0.1

0.15 1

2

3

4

5

6

7

8

9

10

11

12

13

1415

161718

1920

2122

2324

25

26

27

28

29

30

3132 33

34 3536

37

38

39

40

41

42

43

44

45

46

47

48

49

5051

5253

5455

56

5758

5960

61

62

63

64

65

666768 6970 7172

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

8889

90

91

92

93

94

95

9697

98

99100 101

102103

104105106

107108

Single PatientPooled

PC #1 (49%)

PC #

2 (

35

%)

Total Variance


-0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2 0.25

-0.03

-0.02

-0.01

0

0.01

0.02

0.03

0.04

0.05

1

23

4

5

6

7

8

910

11

12 13

1415

1617

18

19

202122

23

24

25

2627

28

29

30

313233

34

35

36

37

383940

41

42

43

44

45

46

47

48 4950

51

525354

555657585960

6162636465

66676869

707172

73 7475767778

79808182

83

84 8586878889 90

919293

94

95

9697

9899100 101102

103104105106

107108

Single PatientPooled

PC #1 (73%)

PC #

2 (

6%

)Treatment vs. Residual


-0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2 0.25

-0.04

-0.03

-0.02

-0.01

0

0.01

0.02

0.03

0.04

0.05

1

2

3

4

5

6

7

8

910

11

12 13

1415

1617

18

19

202122

23

24

25

2627

28

29

30 3132

33

34

35

36

37

383940

41

42

43

44

45

46

47

48 4950

51

525354

555657585960 616263

6465

6667

6869707172

73 747576

7778

79808182

8384 8586878889 90

919293

94

9596

979899

100 101102103

104105106

107108

Day 1Day 2Day 3

PC #1 (76%)

PC #

2 (

3%

)

Day vs. Residual


-0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2 0.25-0.04

-0.03

-0.02

-0.01

0

0.01

0.02

0.03

0.04

0.05

1

23

4

5

6

7

8

910

11

12

13

1415

16

17

18

19

202122

23

24

25

26

2728

29

30

3132333435 36

37

383940

41

42

43

44

45

46

47

48

4950

51

52

5354

55 5657585960 61

6263

6465 66

67686970

7172

7374

7576

7778

79808182

8384

858687888990

919293

9495

9697

98

99

100101

102

103104105106107108

ABC

PC #1 (68%)

PC #

2 (

12

%)

Sample vs. Residual


-0.15 -0.1 -0.05 0 0.05 0.1

-0.2

-0.15

-0.1

-0.05

0

0.05

1

2

3

4

5

6

7

8

9

10

11

1213

14

15

16

17

18

19

20

21

22

23

24

25

26

27

2829

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60 6162

63

64

65

66

67

68

69

70

71

72

73

74

75

76

7778

79

80

81

82

83

84 8586

87

88

89

90

91

92

93

94

95

96 97

98

99100

101

102

103

104

105

106

107

108

ZipTipNothing

PC #1 (68%)

PC #

2 (

23

%)

Center for Intelligent Chemical Instrumentation

• Proteins are captured, retained and purified directly on the chip (affinity capture)

Laser

“Homogeneous” Capture Surface

The SELDI Process and ProteinChipThe SELDI Process and ProteinChip®® Arrays Arrays

• Sample goes directly onto the ProteinChip™ Array

• Array is “read” by Surface-Enhanced Laser Desorption/Ionization (SELDI)

• Retained proteins can be processed directly on the chip

ProteinChipTM Array

Sample

Trace proteins (targets/markers)


ProteinChip® Array Surfaces

Preactivated Surfaces for Specific Protein Interaction Studies

Chromatographic Surfaces for General Profiling

(Reverse Phase) (Cation Exchange) (Metal Ion) (Normal Phase)

(PS-1 or PS-2)(Antibody - Antigen) (Receptor - Ligand) (DNA - Protein)

(Anion Exchange)


ProteinChip® Detection Technology: Laser Desorption Time-of Flight MS

• Retained proteins are detected Laser Desorption Ionization • Simple Linear, TLF TOF MS• Orthogonal Quadrupole TOF MS and MS/MS

Dete

ctor

Dete

ctor

Laser

TOF-MS

0

2.5

5

7.5 Spectra View

8000

2000 4000 6000

2

4

6

2185.8+H

2369.9+H2528.2+H

2781.9+H

3172.3+H

3915.2+H

4000.2+H4345.6+H

4618.2+H

4730.4+H

5045.2+H

7977.1+H

Map View

2000

2000 4000 6000 8000

4000 6000 8000 Gel View


Protein Profiling: Three Dimensions of Resolution

Org

anic

Org

anic

Dete

rgent

Dete

rgent

Salt

Salt

Wate

rW

ate

r

pH

pH

Ure

aU

rea

CH

APS

CH

APS

Imid

azo

lIm

idazo

l

Wash Wash ConditionsConditions

0

2.5

5

7.5

2000 4000 6000 8000

Su

rface

Type

Su

rface

Type

Measured m/zMeasured m/z

12 x 8-spot ProteinChip® Arrays match the footprint of a 96 well microplate


Sample Preparation

• Pre-wash chips with 5% ACN/Methanol

• Deposit 1 μL of sample• Wash chips with 5% ACN• Spot 0.5/1μL of matrix solution ( 3,5-

dimethoxy-4-hydroxycinnamic acid in ACN/H2O/TFA 50/50/0.1)


-1000 -500 0 500

-400

-200

0

200

400

600

1

2

34

5

6

7

8

9

1011

12

13

1415

16

17

18

19

20

21

2223

2425

2627

2829

30

3132

33

34

35

36

373839

4041

42

43

44

45

46

47

48

49

50

51

5253

5455

56

5758

59

60

61

62

63

64

65

66

67

68

69

70

717273

74

75

76

77

7879

80

81

82

83

84

85

86

878889

90

91

9293

94

95

96

97

98

99100101

102

103

104

105106

107

108

109

110

111

112

113114

115

116117

118119

120121

122

123

124

125

126127

128

129130

131132

133

134

135

136137

138

139

140

141142143144145

146147

148

149

150

151

152

153

154

155

156

157

158

159160

161

162163

164165

166167168169

170

171

172

173

174

175176177

178

179

180181

182

183

184

185

186187 188189

190

191

192

193

194

195 196

197

198

199

200

201

202203204205

206

207

208

209

210

211212

213

214

215

216217218

219

220

221

222

223

224

225

03kr7003kr7603kr7803kr7903kr275

PC #1 (59%)

PC #

2 (

12

%)

45 single shot spectra from 5 control serum samples


-400 -200 0 200 400 600 800

-300

-200

-100

0

100

200

1

2

3

4

5

03kr7003kr7603kr7803kr7903kr275

PC #1 (85%)

PC #

2 (

11

%)

Average Spectra


0 2 4 6 8 10

x 104

-0.05

-0.04

-0.03

-0.02

-0.01

0

0.01

0.02

0.03

0.04

0.05

m/z

Rel

ativ

e In

ten

sity

Variable Loadings for the Distribution of the Average Spectra

PC #1PC #2


-1000 -500 0 500 1000

-600

-400

-200

0

200

400

600

1

2

34

5

6

7

8

9

101112

13

1415

16

17

18

19

20

21

2223

2425

2627

282930

3132

33

34

35

36

373839

40

41

42

43

44

45

4647

48

49

50

51

5253

5455

56

575859

6061 62

63

64

65

6667

68

6970

7172

73

74

75

76

77

7879

80

81

82

83

8485

86

878889

9091

9293

9495

96

97

98

99100101

102103

104

105106107

108109

110

111

112

113114

115

116117

118119

120121

122

123

124

125

126127

128

129130

131

132 133134

135

136137

138

139

140

141142143144145

146147

148

149

150

151

152

153

154

155

156

157

158

159160

161

162163164165

166167168169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188189

190

191

192

193

194

195

196

197

198

199

200

201

202

203204205

206

207

208

209

210

211212

213

214

215

216

217218

219

220

221

222

223

224

225

03kr7003kr7603kr7803kr7903kr275

PC #1 (40%)

PC #

2 (

16

%)

Residual Spectra


0 2 4 6 8 10

x 104

-0.05

-0.04

-0.03

-0.02

-0.01

0

0.01

0.02

0.03

m/z

Rel

ativ

e In

ten

sity

Variable Loadings for the Residual Spectra

PC #1PC #2


SELDI Protein Profiles After Depletion of the Highest-abundant Serum Proteins

(albumin, IgG, antitrypsin, IgA, transferrin, haptoglobin) a. Pooled sera from 5 Japanese subjects (01/12/04)b. Individual sera from 5 Caucasian control subjects

(02/19/04)c. Individual sera from 10 Japanese prostate cancer

subjects and 10 matched controls (03/08/04)d. Individual sera from 5 Japanese control subjects,

diluted to 20% and 50% of original concentration and spotted with 1 μL of matrix solution (03/12/04)

e. Individual sera from 3 Japanese control subjects, diluted to 20% and 50% of original concentration and spotted with 0.5 μL of matrix solution


2000 4000 6000 8000 10000

0.005

0.01

0.015

0.02

0.025

0.03

m/z

No

rmal

ized

In

ten

sity


-0.3 -0.2 -0.1 0 0.1 0.2-0.15

-0.1

-0.05

0

0.05

0.1

0.15

1

2

34

5 678

9

10

11

1213

14

15

16

17

18

19

20

21

22

2324

25

26

27

28

29

3031

32

33

34

35

36

37

38

39

40

4142

43

4445

46

47

48

4950

51

52

53

54

55

56

57

58

59

60

6162

63

64

65

66

67

68

69

70

71

72

73

74

7576

7778

79

80

81

82

83

84

85

86

87

8889

90

91

ControlProstate

PC #1 (61%)

PC #

2 (

18

%)


-0.3 -0.2 -0.1 0 0.1 0.2-0.15

-0.1

-0.05

0

0.05

0.1

0.15

1

2

34

5 678

9

10

11

1213

14

15

16

17

18

19

20

21

22

2324

25

26

27

28

29

3031

32

33

34

35

36

37

38

39

40

4142

43

4445

46

47

48

4950

51

52

53

54

55

56

57

58

59

60

6162

63

64

65

66

67

68

69

70

71

72

73

74

7576

7778

79

80

81

82

83

84

85

86

87

8889

90

91

ControlProstate

PC #1 (61%)

PC #

2 (

18

%)

Treatment vs. Residual


Concluding Thoughts• Variability of spectra from MALDI and SELDI

sources are attributable to shot-to-shot variations that are not independent or random.

• Modeling single scans can display chemical and instrumental variations and provide higher quality spectra.

• Mass alignment should be accomplished prior to averaging as opposed to afterwards.

• All the above statements are likely to be attributable to ESI spectra as well.

• PCA coupled to separation of experimental sources of variation provides a useful graphical tool for evaluating experimental procedures.


Acknowledgements

• Students– Libo Cao Matt Rainsberg– George Bota Ping Chen– Preshious Rearden Leyna Denapoli– Leanna Kishler

• Federal Aviation Administration - Donation of a Barringer Ionscan 350

• Ion Track Instruments for Support and Donation of the Itemizer 2 and VaporTracer 1

• Sionex for the donation of DMS– Erkin Nazarov for DMS Slides

• National Biscuit Company - Donation of a GC-MS• U.S. Army EBCB - GeoCenters Donation of 4 Chemical

Agent Monitors and Funding• Research Opportunity Award-Research Corporation • Wright-Patterson Air Force Base-INNSSI Fuel Analysis

center for intelligent chemical instrumentation 1 chemometric considerations in proteomic analyses...

Documents