Constructing high resolution consensus spectra for a peptide library - PowerPoint Presentation

Slide1

Constructing high resolution consensus spectra for a peptide library

Sergey L. Sheetlin,Yuri A. Mirokhin, Dmitrii V. Tchekhovskoi, Xiaoyu Yang, Stephen E. SteinNIST Mass Spectrometry Data Center, Biomolecular Measurement Division,National Institute of Standards and Technology

Slide2

Tandem mass spectrometry

Sample

Ionization

m/z ion sorting

MS1 precursor ion

CID fragmentation

m/z ion sorting

MS2 product ions

Detection

(measuring m/z)

m/z: mass-to-charge ratio

CID: Collision-Induced Dissociation

MS: Mass Spectrum

Slide3

Example of chromatogram (using Thermo

Xcaliber Qual Browser)

Relative abundance

Time (min)

Slide4

Example of MS1 spectrum(using Thermo

Xcaliber Qual Browser)

Relative abundance

Slide5

Example of high resolution MS2 spectrum(shown with NIST MS search program)

Relative abundance

Slide6

Peptides

Short chains of amino acids connected by amide bondsGlu-Thr-Lys

ETK

Glutamylthreonyllysine

C

15

H28N4O7

Slide7

Peptide libraries

NIST MS Search libraries of peptide spectra are available athttp://chemdata.nist.gov/dokuwiki/doku.php?id=peptidew:cdownload

Slide8

Nomenclature for peptide CID fragment ions

R

1

R

2

R

n-1

Rn

| | | | H

2N-CH-CO-NH-CH-CO- …...-NH-CH-CO-NH-CH-CO2

H

I. A. 

Papayannopoulos. The interpretation of collision-induced dissociation mass spectra of peptides. Mass Spectrometry Review, 14:49–73, 1995.

Slide9

Computation of fragment masses

Slide10

Peptides monoisotopic masses

20 amino acids residues and their monoisotopic masses

Ala

Arg

Asn

Asp

Cys

Glu

Gln

Gly

His

Ile

A

R

N

D

C

E

Q

G

H

I

71.0371

156.1011

114.0429

115.0269

103.0092

129.0426

128.0586

57.0215

137.0589

113.0841

Leu

Lys

Met

Phe

Pro

Ser

Thr

Trp

Tyr

Val

L

K

M

F

P

S

T

W

Y

V

113.0841

128.095

131.0405

147.0684

97.0528

87.032

101.0477

186.0793

163.0633

99.0684

Slide11

Modifications

In vivo Posttranslational modifications (PTM) are covalent modifications of proteins after its translation. PTMs play role in activity, function of proteins and their interaction with other molecules.Modifications caused by sample preparation.

Modification name

Composition

Monoisotopic mass

Carbamidomethyl

C

2

H

3

NO

57.021464

iTRAQ4plex

H

12

C

4

13

C

3

N15NO

144.102063

Oxidation

O

15.994915

Deamidation

H

-1

N

-1

O

0.984016

Phosphorylation

HO

3

P

79.966331

Glu

->pyro-

Glu

H

-2

O

-1

-18.010565

Gln

->pyro-

Glu

H

-3

N

-1

-17.026549

Protein modifications for mass spectrometry

www.unimod.org

Slide12

Fragments neutral losses

Common losses are H2O, NH3, CO, H3PO4,iTRAQ (H12C4

13

C

3

N

15NO).

Relative abundance

Slide13

Peptide “GHVIAAR”; charge 2; modification iTRAQ4plex on the first AA ‘G’

Relative abundance

Slide14

Experimental and theoretical isotopic peaks

Relative abundance

Valkenborg

, D.,

Mertens

, I.,

Lemiere

, F.,

Witters

, E.,

Burzykowski

, T.: The isotopic distribution conundrum. Mass

Spectrom

. Rev. 31(1), 96–109 (2012)

Slide15

Annotation of peaks of experimental spectra

Slide16

Experimental and theoretical densities

Probability density function

Slide17

Experimental and theoretical densities

Probability density function

Slide18

Experimental and theoretical densities

Probability density function

Slide19

Error of annotation of experimental peaks

Probability density function

Slide20

Statistics of different types of fragment ions(based on limited set of data)

Slide21

Clustering experimental spectra

Set of unidentified MS2 spectra

Identification (peptide sequencing)

Grouping results by charge, peptide, modifications, collision energy

Filtering

Clusters of replicate spectra

Slide22

Peptide sequencing algorithms

Database search: comparing theoretical spectra for sequences from a database with the query spectrumMS-GF+, Mascot etc.De novo sequencing: trying to find a peptide optimal in terms of some measure of similarity between its theoretical spectrum and the query spectrumPEAKS, NovoHMM etc.Library search: direct comparison of the query spectrum with identified spectra from a library

NIST Mass Spectral Library etc.

Slide23

Example of experimental spectra from the same cluster

Relative abundance

Slide24

Computing peaks of consensus spectra

Replicate spectra

Consensus

 

 

 

 

 

 

 

 

MS-GF+

Slide25

Computing peaks of consensus spectra

Slide26

Example of log-likelihood

Slide27

Average fraction of replicates

Hypothesis: “good” peaks of the consensus spectrum have properties similar to annotated peaks

Slide28

Filtering peaks of consensus spectra

Replicate number

Is there a peak in the replicate

corresponding to the

given consensus peak with abundance A?

1

Yes

2

Yes

…

…

N-1

No

N

Yes

Bernoulli distribution:

Yes

with probability p;

No

with probability 1-p

Slide29

Density of peaks for different relative abundancies

Slide30

Comparison of consensus and best replicate spectra

Slide31

Further directions

Adjusting the parameters of the method for optimal performance of the existing search algorithmsBuilding peptide libraries of consensus spectra

Slide32

Acknowledgements

NIST MS Data CenterYuri A. MirokhinDmitrii V. Tchekhovskoi

Xiaoyu Yang

Stephen E. Stein

William E. Wallace