Cheminformatics and mass spectrometry course

Download Report

Transcript Cheminformatics and mass spectrometry course

Welcome!
Mass Spectrometry meets Cheminformatics
Tobias Kind and Julie Leary
UC Davis
Course 4: Mass Spectrometry Tools & Concepts
Class website: CHE 241 - Spring 2008 - CRN 16583
Slides: http://fiehnlab.ucdavis.edu/staff/kind/Teaching/
PPT is hyperlinked – please change to Slide Show Mode
1
Atomic Mass
Correct unit is [u] – unified atomic mass unit or [Da] Dalton see SI units
1 u = 1 Da = 1/12th of mass of carbon 12C = 1.66053886 x 10-27 kg
C6Cl6: C6 Cl6 p(gss, s/p:40) Chrg 0R: 1000 Res.Pwr...
283.81
100
90
Hexachlorobenzene (C6Cl6)
Cl
285.81
80
Cl
Cl
Cl
Cl
Relative Abundance
70
60
281.81
50
Cl
average mass
-
284.7804 u
integer mass
-
282.0 u
40
287.80
monoisotopic mass -
30
281.81312 u
20
282.82
284.81
10
286.81
289.80
288.81
291.80 292.80 294.80 295.80
0
282
284
286
288
290
292
294
296
m/z
Always (always) check molecular masses obtained from databases or publications.
For mass spectrometry the monoisotopic mass is used.
InChIKey:CKAPSXZOOQJIBF-UHFFFAOYAV
2
MWTWIN
Example: Molecular Weight Calculator (Matthew Monroe / PNNL)
3
Mass Accuracy
Instruments must be calibrated to obtain high mass accuracy.
In case of FT-ICR-MS mass calibration can be stable over weeks.
Post- mass calibration can be performed if calibrant was run with samples.
Mass of electron becomes important at around 500 Da.
Type
Mass Accuracy
FT-ICR-MS
0.1 - 1 ppm
Orbitrap
0.5 - 1 ppm
Magnetic Sector
1 - 2 ppm
TOF-MS
3 - 5 ppm
Q-TOF
3 - 5 ppm
Triple Quad
3 - 5 ppm
Linear IonTrap
50-200 ppm
(10 ppm in Ultra-Zoom)
ppm  (
mexp - mcalc
)  1 E 6
mexp
m(e-)
= 0.00054858026 u = mass of electron
m(1H)
= 1.0078246 u = mass of proton
4
Resolving Power
RP = 1700
High resolving power is helpful for
separation of species with almost
same mass (isobars).
RP = 48,250
High resolving power can not be
used to distinguish between
structural isomers.
Example:
C8H10N2O has 100,082,479 isomers.
5
Example Solanine (CID=30185)
Isotopic Pattern Generators
Elements can be
a) monoisotopic (F, Na, P, I)
b) polyisotopic (H, C, N, O, S, Cl, Br)
Isotopic pattern generators generate the isotopic abundances for a given mass value.
Calculation is very time-consuming and based on Fast Fourier algorithms.
100.0
80.0
60.0
40.0
20.0
Example: Molecular Weight Calculator (Matthew Monroe / PNNL)
0.0
866.00
868.00
870.00
872.00
874.00
876.00
6
878.00
Isotopic pattern generators
C6Cl6: C6 Cl6 p(gss, s/p:40) Chrg 0R: 1000 Res.Pwr...
283.81
100
90
285.81
80
Relative Abundance
70
60
281.81
50
40
287.80
30
20
282.82
284.81
10
286.81
289.80
288.81
291.80 292.80 294.80 295.80
0
282
284
286
288
290
292
294
296
m/z
Example form Thermo Xcalibur with a very versatile isotopic pattern generator
7
Charge states
charge state 1
charge state 2
CID: 3081765
MW = 1125.50082
C50H72N13O15P
8
Different charge states and peak resolutions
562.75
100
100
1125.50
90
90
2000 Resolving Power
Charge state 2
80
70
563.25
60
2000 Resolving Power
Charge state 1
80
70
1126.51
60
C 50 H 72 N 13 O 15 P:
C 50 H 72 N 13 O 15 P 1
p (gss, s /p:40) Chrg 2
R: 2000 Res .Pwr . @FWHM
50
40
50
C 50 H 72 N 13 O 15 P:
C 50 H 72 N 13 O 15 P 1
p (gss, s /p:40) Chrg 1
R: 2000 Res .Pwr . @FWHM
40
30
30
563.76
20
10
564.26
0
100
10
564.76 565.77 566.78
567.78
562.75
0
100
90
90
80
80
200,000 Resolving Power
Charge state 2
70
60
563.25
50
0.5
40
30
20
563.75
10
564.25
0
562
563
564
1127.52
20
C 50 H 72 N 13 O 15 P:
C 50 H 72 N 13 O 15 P 1
p (gss, s /p:40) Chrg 2
R: 200000 Res .Pwr . @FWHM
1128.52
1130.54 1132.55 1134.55
1125.50
60
1126.50
50
1.0
40
30
20
1127.51
10
564.76 565.76 566.76 567.26
565
m/z
566
567
568
200,000 Resolving Power
Charge state 1
70
1128.51
0
1125
Example of Phosphorylated Angiotensin isotopic pattern without adduct
C 50 H 72 N 13 O 15 P:
C 50 H 72 N 13 O 15 P 1
p (gss, s /p:40) Chrg 1
R: 200000 Res .Pwr . @FWHM
1129.51 1131.52 1133.52 1135.53
1130
m/z
[M+H]+
1135
simulated by Thermo XCalibur
9
Charge state deconvolution
z=2
Monoisotopic peaks are shown on
the mass spectrum, where z = 1
z=1
Example from Decon2Ls Tutorial (PNNL)
A Software Tool for Deconvolution of
High Resolution Mass Spectra
Peaks are detected and charge states are
automatically calculated for whole datasets
Download Decon2LS
10
Picture sources: PNNL/ Decon2Ls Tutorial
Adduct formation
Adduct formation is observed in ESI, CI and APCI ionization modes (and others).
Adduct detection is problematic for small molecules, can be influenced by solvent selection.
Adduct detection can be automated if two or more adducts are detected in mass spectrum.
Switching polarity (+/-) can be used for confirmation of adduct.
Ion name
1. Positive ion mode
M+H
M+NH4
M+Na
M+CH3OH+H
1. Negative ion mode
M-H
M+Na-2H
M+Cl
Ion mass
Charge Mult Mass
Result:
M + 1.007276
M + 18.033823
M + 22.989218
M + 33.033489
1+
1+
1+
1+
1
1
1
1
1.007276
18.033823
22.989218
33.033489
2.007276
19.033823
23.989218
34.033489
M - 1.007276
M + 20.974666
M + 34.969402
111-
1
1
1
-1.007276
20.974666
34.969402
-0.007276
21.974666
35.969402
Download the Mass Spectral Adduct Calculator
11
http://fiehnlab.ucdavis.edu/staff/kind/Metabolomics/MS-Adduct-Calculator/
Molecular Formula Generators
Formula generators are used to create molecular formulae from accurate masses.
Input requires
1) accurate isotopic mass (with or without adduct) and
2) error in ppm or mDa (milli Dalton)
Accurate mass
Mass error
Example MWTWIN
12
The molecular formula space of small molecules
calculated by the Seven Golden Rules
Each molecular formula can expand to billions of structural isomers.
Molecular Formula ≠ Molecular Isomer
13
Source http://fiehnlab.ucdavis.edu/projects/Seven_Golden_Rules/Molecular-Formula-Space/
Frequency distribution of molecular formulas
14
Source: Kind and Fiehn BMC Bioinformatics 2006 7:234 doi:10.1186/1471-2105-7-234
Impact of mass accuracy on number of formulas
15
Mass accuracy and isotopic pattern
[M+H]+
C45H73NO15
MW = 867.49799
Example:
ESI-MS (+) of Solanine on a LTQ
Resolving Power: 1700
Mass Accuracy: 46 ppm
Isotopic Abundance Error: ±1.46%
16
Example from: http://fiehnlab.ucdavis.edu/projects/Seven_Golden_Rules/Examples/Solanine/
Isotopic abundances as orthogonal filter
17
Source: Kind and Fiehn BMC Bioinformatics 2006 7:234 doi:10.1186/1471-2105-7-234
Molecular Isomer Generators
Isomer generators are used to create all possible structural isomers from a
given molecular formula. Deterministic and stochastic (random) generators are in use.
Example: MOLGEN DEMO (Bayreuth)
18
The molecular isomer space is unknown
Accurate mass
Formula
Number Isomers
in Beilstein DB
77.99531
CH2O4
6
0
78.04293
CH6N2O2
28
1
78.03169
C2H6O3
10
8
78.02180
C4H2N2
465
2
78.01056
C5H2O
151
2
78.04695
C6H6
217
29
150.04293
C7H6N2O2
100,082,479
153
150.09054
C7H10N4
66,583,863
105
150.03169
C8H6O3
6,717,404
90
150.07931
C8H10N2O
76,307,072
542
150.06808
C9H10O2
6,843,602
667
150.11569
C9H14N2
9,459,132
568
150.02180
C10H2N2
65,563,828
0
150.10446
C10H14O
1,548,361
1938
150.01056
C11H2O
9,414,509
0
150.14084
C11H18
84,051
762
150.04695
C12H6
34,030,905
12
Source: http://fiehnlab.ucdavis.edu/projects/Seven_Golden_Rules/Molecular-Isomer-Generator/
Meringer M: Mathematische Modelle für die kombinatorische Chemie und die molekulare Strukturaufklärung.
19
MOLGEN-MS
20
Substructure predictions
Use computer algorithms (machine learning) for automated interpretation of fragments
and corresponding substructures. Algorithms creates present/absent list of substructures.
300
100
50
150
0
100
168 222
40
130
(mainlib) Coronene
220
310
21
Example from NIST-MS search
Simulation of mass spectra
Why is simulation or prediction of mass spectra important?
 Molecular isomers (structures) can be generated very fast from molecular formulas
 Only certain mass spectra can be simulated (MS/MS of peptides, oligosaccharides, lipids)
 Problematic is abundance determination
 Problematic are all complex rearrangement reactions (gas phase ion chemistry)
 Simulation of mass spectra from small molecules is new and important research
Experimental
spectra
Simulated
spectra
Perform comparison or matching
22
Barcode spectrum from Mass Frontier
Example: MassFrontier Helpfile
23
Peptide Sequence Fragmentation Modelling
24
Example: Taken from Molecular Weight Calculator (Matthew Monroe / PNNL)
Matching experimental vs. theoretical sequence
Example: RHPEYAVSVLLR
25
The Last Page - What is important to remember:
Important performance parameters are accurate mass, resolving power, scan speed
and accurate isotopic abundances
Even high resolving power MS can not distinguish between structural isomers
Accurate Mass  Molecular Formula  Structural Isomers  MS/MS
Mass spectra of only some substance classes can be simulated
26
Tasks (12 min):
(1)
Use a Molecular Formula Finder
Molecular Weight Calculator or any other tool (HR2 in 7 Golden Rules)
Calculate how many formulae (elements CHNO) can be calculated for
a) m/z 500.5 at 100 ppm, 10 ppm, 1 pm
b) m/z 500.0 at 100 ppm, 10 ppm, 1 pm
(2)
Calculate the number of isomers for C12H12 using MOLGEN or CDK
(3)
Generate the isotopic pattern for Chlorophyll a and Hexachlorobenzene.
(4)
Download Decon2LS and perform example calculations from tutorial
27
Literature (35 min):
Debating resolution and mass accuracy (MP Balogh)
Enumerating molecules (Faulon)
Milestones in fourier transform ion cyclotron resonance mass spectrometry technique development
28
Links:
Used for research: (right click – open hyperlink)
http://download.enovatia.com/images/promass/help/promassxcaliWebMain.html
http://download.enovatia.com/images/promass/protein_MW_determination/bsa.html#1.36zoom
http://www.bioc.aecom.yu.edu/labs/angellab/
http://www.ads.tuwien.ac.at/publications/bib/pdf/mujezinovic_07.pdf
http://mendel.imp.ac.at/mass-spectrometry/
http://www.chm.bris.ac.uk/sillymolecules/sillymols.htm
http://ncrr.pnl.gov/training/workshops/2007HUPO/LCMSBasedProteomicsDataProcessing.pdf
http://www.iupac.org/publications/pac/2003/pdf/7506x0683.pdf
"Sic transit gloria chimica”
http://www.google.com/search?hl=en&q=%22Sic+transit+gloria+chimica&btnG=Search
Of general importance for this course:
http://fiehnlab.ucdavis.edu/staff/kind/Metabolomics/Structure_Elucidation/
29