high throughput workflow for global metabolomic profiling ... · high throughput workflow for...
TRANSCRIPT
High Throughput Workflow for Global Metabolomic Profiling with RapidFire online
SPE/MS system
Michelle Romm, Ph.D.Applications Scientist,
Agilent Technologies, Inc.
6/23/2015
1
What is High Throughput?
6/23/2015
2
# of Samples
N=1
N=100
N=1,000
N=10,000
LC/MS
30 mins
50 hrs
20.8 days
208.3 days
RapidFire/MS
14 secs
23 mins
3.8 hrs
38.8 hrs (1.6 days)
120X Faster
Why High Throughput Profiling?
6/23/2015
3
Method Development Fast Turnaround Large Population StudiespH
Solvents/AdditivesChemistry
Timings/Flows
N= 5505005000
QC/QAClass Prediction
Is Faster Better?When
6/23/2015
4
RapidFire/MS System
RapidFire 365 Online SPE/MS System
For Research Use Only. Not for Use in Diagnostic Procedures.
6/23/2015
5
BenchBot plate handling robot63 plate capacity (384 well plates = 24,192
samples)Off-deck reach capability
Quaternary pumpsSolvent switching capability Capable of running multiple disparate methods in
one batch
12 position automated cartridge changer Automated method development
Perform multiple injections of a single sample with different SPE cartridges and mobile phasesC18, C8, C4, CN, Phenyl, HILIC, Graphitic
Carbon Simply run multiple samples overnight and
select your ideal conditions the next morning
Aspirate Load/Wash Elute Re-equilibrate
600ms 3000ms 8000ms 1000ms
Method Development: Refining the Experiments
• Cartridge chemistries
• Ionization mode
• Loading solvents
• Elution solvents
One solution doesn’t fit everyone, this system allows FLEXIBILITY!
6/23/2015
6
Comparison of Chemistries: C18 vs. HILIC SPE/MS (+ESI) in Urine
526 metabolites annotated by C18 SPE/MS
270 metabolites annotated by HILIC SPE/MS
49 annotated metabolites common between HILIC and C18SPE/MS
6/23/2015
7
HILIC221
C18477
49
Utilize multiple unique chemistries to increase the metabolite coverage
747 unique metabolites (100% reproducibility; CV<15%)
Comparison of ESI Mode: a Small Polar Library
6/23/2015
8
Red- NEGBlue- POS
• ~100 pure endogeneous human metabolite standards• small organic acids, biogenic amines, nucleotides, amino acids
• Custom packed HILIC material
Comparison of Loading Solvents: HILIC Normal Phase
6/23/2015
9
ACN:Water (95:5)
Hexane:IPA (90:10)
Methanol
Lactic Acid, -ESI
Comparison of Elution pH
6/23/2015
10
MeOH:IPA (50:50) +0.1% Formic Acid
MeOH:IPA (50:50) +10mM Ammonium Acetate pH 9.0Lactic Acid, -ESI
High Throughput Metabolomics Profiling WorkflowBatch Data Processing Data Analysis
Verification/ Confirmation/
Validation
High ThroughputQuantitation
High ThroughputAcquisition
MassHunter Profinder
(NEW)
.d data input
RecursiveMFE
Features.cef output
Mass ProfilerProfessional
(MPP)
Statistics
Database Search
Pathway Analysis
Potential Targets
Pooled Sample/Random Samples
LC/MSAcquisitionFindbyFormulaVerify Targets
LC/MS/MSAcquisition
Metlin Library/SimLipid/
MSC
Confirm Targets
Validate withReference Standard
RapidFire 365
iFunnel 6550 Q-TOF
RapidFire 365
iFunnel 6490 QQQ
6/23/2015
11
Human Urine Metabolomics Profiling: Sample Prep Human urine, collected and pooled from healthy subjects (males and females), was purchased
from Innovative Research Inc. (Novi, MI)
Filter 400µL urine through a 0.22µm Ultrafree-MC® centrifugal filter units (Millipore) bycentrifugation @ 12,000xg for 20min @ 4°C
Pipette 50µL of filtered urine (10 replicates) into the Amicon Ultra-0.5mL Centrifugal Filter (3kDaCut-off Ultracel-3 membrane) from Millipore
Dilute with 350µL of Milli-Q quality water containing 0.2% Acetic acid, 1mM Ammonium acetate,and centrifuge @ 12000xg, 4°C for 30min to concentrate
The flow-through represents the metabolomics sample
Inject 10µL of flow-through for SPE/MS as well as LC/MS/MS analysis
Human urine equivalent to 1.25µL was used for the metabolomics profiling studies
6/23/2015
12
TIC for 10 Human Urine Samples (C18 SPE/MS)
6/23/2015
13
Single Injection of Urine Sampleon C18 RapidFire SPE/MS
6/23/2015
14
6x10
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
1.9
2
2.1
2.2
2.3
2.4
Counts vs. Acquisition Time (min)
0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0.11 0.12 0.13 0.14 0.15 0.16 0.17 0.18 0.19 0.2 0.21 0.22 0.23
Male Urine+ESI
A Single Feature in Urine on C18 RapidFire SPE/MS
6/23/2015
15
4x10
0
0.25
0.5
0.75
1
1.25
1.5
1.75
2
2.25
2.5
2.75384.0461 (M+H)+
385.0466 (M+H)+
386.0505 (M+H)+406.0273 (M+Na)+
Counts vs. Mass-to-Charge (m/z)
379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412
Male Urine
5x10
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.1
1.2
1.3
1.4
0.147 Cpd 323: 0.147
Counts vs. Acquisition Time (min)
0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0.11 0.12 0.13 0.14 0.15 0.16 0.17 0.18 0.19 0.2 0.21 0.22 0.23 0.24 0.25 0.26 0.27 0.28 0.29
EIC (C18, Pos ESI)Male Urine
ProFinder for High Throughput Batch Analysis of Samples
6/23/2015
16
Mass Profiler Professional (MPP)
6/23/2015
17
Cluster Analysis Filter by all 10 reps, abundance >2, CV < 15%
C18-PosESI
Annotation of Creatinine in Urine Samples by C18 RapidFire SPE/MS
6/23/2015
18
Select High Abundant Metabolites in Human Urine Annotated by SPE/MS
6/23/2015
19
Compound RF-C18 MaleHippuric acid 5158031Pirimicarb 3255466Enprofylline 496924Lys Asn Asp 4201013-Hydroxyhippuric acid 410564Pinacidil 359491Nadolol 3572284-(1-Hydroxy-2-(methylamino)ethyl)phenol 299538Leu Gln Lys 287646Triciribine 2859323-Amino-2-naphthoic acid 277964L-4-Hydroxy-3-methoxy-a-methylphenylalanine 2660331-Methylxanthine 262927AFMK 229545Creatinine 2252744-Chloroacetophenone 217597Jasmolone glucoside 209979Istamycin C1 192524N-Acryloylglycine 183487N2,N2-Dimethylguanosine 175089N(alpha)-t-Butoxycarbonyl-L-leucine 171812beta-Snyderol 1351772-Methylguanosine 126988
6/23/2015
20
Metlin Library Check for Isomers/Isobars
LC/MS/MS Confirmation of Hippuric acid
MFG for MS/MS Spectral PeaksHippuric acid
CID@10
CID@20
CID@40
Hippuric acidMS/MS @ 40Metlin Spectral Library
RapidFire C18 SPE/MS +ESI
6/23/2015
21
Male Urine
6/23/2015
22
300 newborn dried blood spot cards (males and females), were obtained fromthe state of California
3.2mm punch was taken from center of each spot and extracted in 200uL of icecold 80:20 Methanol:Water containing internal standards followed bycentrifugation @ 12,000xg for 10min @ 4°C
Supernatant was passed through a Phenomenex Phree SPE plate to removephospholipids
200uL of 20:80 Methanol:Water was added to flow through
10uL of final solution (equivalent to ~1.5uL of blood) was analyzed on the AgilentRapidFire in-line SPE system using Imtakt Scherzo SM-C18 mixed mode andbare silica HILIC cartridges in positive and negative ionization modes
Dried Blood Spot Metabolite Profiling: Sample Prep
6/23/2015
23
12
14
16
18
20
22
2413C Glutamine%CV = 7.6
13C Glucose%CV = 8.2
300 blood spots
Sig
nal I
nten
sity
What is the variability? Internal Standards
6/23/2015
24
10
12
14
16
18
20
22
Glutamate
10
11
12
13
14
15
16
17
Methionine
10
11
12
13
14
15
16
17
18
19
Histidine
10
12
14
16
18
20
22
24
Free Carnitine
10
11
12
13
14
15
16
17
18
19
20
Propionylcarnitine
10
11
12
13
14
15
16
17
18
Hexanoylcarnitine
Metabolite Verification
6/23/2015
25
PCA Analysis of DBS: Can we see differences?
Present in 75%, not present in solvent or blank blood card, abundance >10,000, fold change >2
SMC18-PosESI
6/23/2015
26
Gala Fuji Gold Delicious
Granny Smith
Red Delicious
75 apples total, 5 apples of each variety from 3 different stores
Peel was carefully collected from each apple with care to avoid removing any of theapple flesh or harvest from damaged areas
Peels were cryomilled, lyophilized and then extracted in methanol at -20C for 12 hours
Inject 10µL of sample for SPE/MS
Class Prediction Analysis: Sample Prep
6/23/2015
27
6x10
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.1
1.2
1.3
1.4
1.5
1.6
1.7
Cpd 1: 0.128: +ESI ECC Scan Frag=175.0V Inj00008-Apple Peel-C18-A3
0.128 Cpd 1: 0.128
Counts vs. Acquisition Time (min)
0.11 0.12 0.13 0.14 0.15 0.16
7x10
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
+ESI TIC Scan Frag=175.0V Inj00008-Apple Peel-C18-A3-F-9.d
Counts vs. Acquisition Time (min)
0.11 0.12 0.13 0.14 0.15 0.16
6x10
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
2.2
Cpd 1: 0.121: +ESI ECC Scan Frag=175.0V Inj00010-Apple Peel-GC-A5
0.121 Cpd 1: 0.121
Counts vs. Acquisition Time (min)
0.11 0.12 0.13 0.14 0.15
7x10
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.1
1.2
1.3
1.4
1.5
+ESI TIC Scan Frag=175.0V Inj00010-Apple Peel-GC-A5-GS-13.d
Counts vs. Acquisition Time (min)
0.11 0.12 0.13 0.14 0.15
6x10
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
2.2
2.4
2.6
2.8
Cpd 1: 0.120: +ESI ECC Scan Frag=175.0V Inj00011-Apple Peel-E9-RD-
0.120 Cpd 1: 0.120
Counts vs. Acquisition Time (min)
0.11 0.12 0.13 0.14 0.15 0.16
7x10
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.1
+ESI TIC Scan Frag=175.0V Inj00011-Apple Peel-E9-RD-10.d
Counts vs. Acquisition Time (min)
0.11 0.12 0.13 0.14 0.15 0.16
C18 Fuji
Graphitic CarbonGranny Smith
HILICRed Delicious
TIC 1
ECC 1
TIC 2 TIC 3
ECC 2 ECC 3
Different Cartridge Profiles on RapidFire SPE/MS
6/23/2015
28
ProFinder: Batch Analysis
6/23/2015
29
Features Detected
Variety C18Graphitic Carbon HILIC
Fuji 1017 404 229
Gala 991 411 227
Golden Delicious 987 411 227
Red Delicious 1005 421 229
Granny Smith 996 419 228
Data Analysis Workflow: Summary
Frequency 100% in one group, not present in blank, One-Way ANOVA (p<0.05, Benjamini-Hochberg corr)
Which features distinguish between the apple varieties?
6/23/2015
30
Partial Least Squares Discriminate Analysis Model for C18
R2 = 0.85Q2 = 0.57
Fuji
Gala
Gold D
Granny S
Red D
6/23/2015
31
R2 = 0.75Q2 = 0.35
Fuji
Gala
Gold D
Granny S
Red D
Partial Least Squares Discriminate Analysis Model for HILIC
6/23/2015
32
Model Building Workflow: Prediction Results
6/23/2015
33
RapidFire - Ion Mobility QTOF
System sensitivity optimized using electrodynamic ion funnels to focus and transmit ions
Ion Mobility resolution optimized while maintaining QTOF performance (mass resolution and accuracy)
Ion Fragmentation can be selected using standard QTOF collision cell (CID)
Bandwidth of QTOF data acquisition and processing channel was increased by 10 fold to accommodate the ion mobility data
IM-QTOF Instrument Overview
34
6/23/2015
35
Wash Blank
Trisaccharides
Melezitose
Raffinose
Why Ion Mobility: Separation!Resolving Structural Sugar Isomers C18H32O16
Tiabendazole
Imazaquin
Why RapidFire-IMQTOF: Sensitivity!
6/23/2015
37
Metabolites extracted from mice plasma
Metabolites extracted from human urine
Isomeric Separation: m/z=330.228
Same nominal mass: m/z=138.054, 138.129
Isomeric Separation: m/z=663.485
Analysis of Metabolomes from Different Matrices
Take Home Message…
6/23/2015
38
Using RapidFire/MS for metabolomics profiling provides a high throughput workflow that is flexible, efficient,
reproducible and produces high fidelity data
Agilent Technologies• Dan Cuthbertson• Sumit Shah
UC San Diego• Jeramie Watrous• Mohit Jain
PNNL• Xing Zhang• Erin Baker
6/23/2015
39
Acknowledgements