Mass Spectrometrybased Proteomics Xuehua Shen Adapted from slides

Mass Spectrometry-based Proteomics Xuehua Shen (Adapted from slides with textbook)

Outline • Motivation of proteomics • Mass spectrometry-based proteomics • Instrumentation of mass spectrometry • De novo sequencing algorithm • Database search • Algorithms of real software (e. g. , sequence tags) 2

Motivation • Proteins are working units of the cells – The number of found genes is much less than the number of expressed proteins – Directly related with cell processes and diseases DNA SNP ~30, 000 human genes m. RNA Protein Alternative Post-translational splicing Modification >100, 000 RNA messages >1, 000 distinct protein forms 3

Tools for Proteomics • Edman degradation reaction • NMR (Nuclear Magnetic Resonance) • X-ray crystallography • Protein array • Mass Spectrometry 4

Mass Spectrometry-based Proteomics • Primary sequence (sequencing, identification) • Post-translational modification (PTM) (characterization) • Quantitative proteomics (quantification) • Protein-protein interaction 5

6

Components of Mass Spectrometer • Ion source (ESI and MALDI) • Mass analyzer (ion traps, TOF, Quadrupole, FT, etc. ) – Mass-to-charge ratio (m/z) • Ion detector 7

Peptide and Intact Protein • Peptide: a fragment of protein • Some enzymes, e. g. trypsin, break protein into peptides. • Some technology put intact protein into the mass spectrometer 8

Peptide Fragmentation Collision Induced Dissociation H+ H. . . -HN-CH-CO Ri-1 N-Terminus • • . . . NH-CH-CO-…OH Ri Ri+1 C-Terminus Peptides tend to fragment along the backbone. Fragments can also loose neutral chemical groups like NH 3 and H 2 O. 9

Ideal Mass Spectrum 10

Real Mass Spectrum 11

rm te C- es tid ep es id pt pe lp ina al in rm te N- N- and C-terminal Peptides 12

Terminal peptides and ion types Peptide Mass (D) 57 + 97 + 147 + 114 = 415 without 57 + 97 + 147 + 114 – 18 = 397 13

N- and C-terminal Peptides 4 86 71 ep pt 57 185 lp pe ina al 332 C- te rm in rm te N- 154 tid id 30 1 es es 415 429 14

N- and C-terminal Peptides 4 86 71 ep pt 57 185 lp pe ina al 332 C- te rm in rm te N- 154 tid id 30 1 es es 415 429 15

N- and C-terminal Peptides 4 86 415 71 30 1 185 154 332 57 429 16

N- and C-terminal Peptides 4 86 71 415 30 1 154 57 Problem: Reconstruct peptide from the set of masses of fragment 185 332 429 17

Mass Spectra 57 Da =K‘G’ D D V 99 Da = ‘V’ L H 2 O G L D K V G mass 0 • The peaks in the mass spectrum: – Prefix and Suffix Fragments. – Fragments with neutral losses (-H 2 O, -NH 3) – Noise and missing peaks. 18

Protein Identification with MS/MS G V D K Peptide Identification: Intensity MS/MS L mass 00 19

Protein Identification by Tandem Mass Spectrometry S e q u e n c e MS/MS instrument De Novo interpretation • Sherenga Database search • Sequest 20

De Novo vs. Database Search De Novo Mass, Score W Database of known peptides V A A MDERHILNM, KLQWVCSDL, PTYWASDL, ENQIKRSACVM, TLACHGGEM, NGALPQWRT, HLLERTKMNVV, GGPASSDA, GGLITGMQSD, MQPLMNWE, ALKIIMNVRT, AVGELTK, , HEWAILF, GHNLWAMNAC, GVFGSVLRA, EKLNKAATYIN. . C L G G P R L L E T K K W D T AVGELTK 21

Pros and Cons of de novo Sequencing • Advantage: – Gets the sequences that are not necessarily in the database. • – An additional similarity search step using these sequences may identify the related proteins in the database. Disadvantage: – Requires higher quality data. – Often contains errors. 22

Current Status • It is still a open problem of protein sequencing no matter whether using de novo sequencing or database search methods • Following algorithms only deal with simplified (or ideal) spectrums • Some algorithms combine de novo sequencing and database search 23

Outline • Motivation of proteomics • Mass spectrometry-based proteomics • Instrumentation of mass spectrometry • De novo sequencing • Database search • Algorithms of real software (e. g. , sequence tags) 24

De novo Peptide Sequencing Sequence 25

Peptide Sequencing Problem Goal: Find a peptide with maximal match between an experimental and theoretical spectrum. Input: – S: experimental spectrum – Δ: set of possible ion types – m: parent mass Output: – P: peptide with mass m, whose theoretical spectrum matches the experimental S spectrum the best 26

Procedure of De Novo Sequencing • Build spectrum graph – How to create vertices (from masses) – How to create edges (from mass differences) • Find best path or rank paths of spectrum graph – How to find candidate paths – How to score paths 27

From Sequence to Spectrum b S E Q U E N Mass/Charge (M/Z) C E 28

From Sequence to Spectrum (cont. ) a SE Q U E N Mass/Charge (M/Z) C E 29

From Sequence to Spectrum (cont. ) a is an ion type shift in b S E Q U E Mass/Charge (M/Z) N C E 30

From Sequence to Spectrum (cont. ) y E C N E U Q Mass/Charge (M/Z) E S 31

Intensity From Sequence to Spectrum (cont. ) Mass/Charge (M/Z) 32

Intensity From Sequence to Spectrum (cont. ) Mass/Charge (M/Z) 33

From Sequence to Spectrum (cont. ) noise Mass/Charge (M/Z) 34

Intensity MS/MS Spectrum Mass/Charge (M/z) 35

Some Mass Differences between Peaks Correspond to Amino Acids u q s e e c e u q e n n q u e n c c e e s e 36

Now decoding from spectrum to sequence…? Build spectrum graph 37

Vertices of Spectrum Graph • Vertices are generated by reverse shifts corresponding to ion types • Δ={δ 1, δ 2, …, δk} Every mass s in an MS/MS spectrum generates k vertices V(s) = {s+δ 1, s+δ 2, …, s+δk} corresponding to potential N-terminal peptides • Vertices of the spectrum graph: {initial vertex} V(s 1) V(s 2) . . . V(sm) {terminal vertex} 38

Reverse Shifts Shift in H 2 O+NH 3 39

Edges of Spectrum Graph • Two vertices with mass difference corresponding to an amino acid A: – Connect with an edge labeled by A • Gap edges for di- and tri-peptides – Potential sequence tag method (covered later) 40

Best Path of Spectrum Graph • How to find candidate paths • There are many paths, how to find the correct one? • We need scoring to evaluate paths 41

Find Candidate Paths • Heuristics: find a path with maximum number of edges • Longest path problem in DAG • DFS (Depth First Search) 42

Path Score • p(P, S) = probability that peptide P produces spectrum S= {s 1, s 2, …sq} • p(P, s) = the probability that peptide P generates a peak s • Scoring = computing probabilities 43

Finding Optimal Paths in the Spectrum Graph • For a given MS/MS spectrum S, find a peptide P’ maximizing p(P, S) over all possible peptides P: • Peptides = paths in the spectrum graph • P’ = the optimal path in the spectrum graph • Some software rank paths 44

Ions and Probabilities • A peptide has all k peaks with probability • and no peaks with probability • A peptide also produces a ``random noise'' with uniform probability q. R in any position. 45

Ratio Test Scoring for Partial Peptides • Incorporates premiums for observed ions and penalties for missing ions. • Example: for k=4, assume that for a partial peptide P’ we only see ions δ 1, δ 2, δ 4. The score is calculated as: 46

Why Not Sequence De Novo? • De novo sequencing is still not very accurate! Amino Acid Accuracy Whole Peptide Accuracy 0. 566 0. 189 SHERENGA (Dancik et. al. , 1999). 0. 690 0. 289 Peaks 0. 673 0. 727 0. 246 0. 296 Algorithm Lutefisk (Taylor and Johnson, 1997). (Ma et al. , 2003). Pep. Novo (Frank and Pevzner, 2005). • Less than 30% of the peptides sequenced were completely correct! 47

The End Thank you ! 48

De Novo vs. Database Search De Novo W Database of known peptides V A A MDERHILNM, KLQWVCSDL, PTYWASDL, ENQIKRSACVM, TLACHGGEM, NGALPQWRT, HLLERTKMNVV, GGPASSDA, GGLITGMQSD, MQPLMNWE, ALKIIMNVRT, AVGELTK, , HEWAILF, GHNLWAMNAC, GVFGSVLRA, EKLNKAATYIN. . C L G G P R L L E T K K W D T AVGELTK 49

De Novo vs. Database Search: A Paradox • • • de novo algorithms are much faster, even though their search space is much larger! A database search scans all peptides in the search space to find best one. De novo eliminates the need to scan all peptides by modeling the problem as a graph search. Why not sequence de novo? 50

Outline • Motivation of proteomics • Mass spectrometry-based proteomics • Instrumentation: Mass Spectrometry • De novo sequencing algorithm • Database search • Algorithms of real software (e. g. , sequence tags) 51

Peptide Identification Problem Goal: Find a peptide from the database with maximal match between an experimental and theoretical spectrum. Input: – S: experimental spectrum – database of peptides – Δ: set of possible ion types – m: parent mass Output: – A peptide of mass m from the database whose theoretical spectrum matches the experimental S spectrum the best 52

MS/MS Database Search Database search in mass-spectrometry has been very successful in identification of already known proteins. Experimental spectrum can be compared with theoretical spectra of database peptides to find the best fit. SEQUEST (Yates et al. , 1995) But reliable algorithms for identification of modified peptides is a much more difficult problem. 53

Post-Translational Modifications Proteins are involved in cellular signaling and metabolic regulation. They are subject to a large number of biological modifications. Almost all protein sequences are posttranslationally modified and 200 types of modifications of amino acid residues are known. 54

Examples of Post-Translational Modification Post-translational modifications increase the number of “letters” in amino acid alphabet and lead to a combinatorial explosion in both database search and de novo approaches. 55

Search for Modified Peptides: Virtual Database Approach Yates et al. , 1995: an exhaustive search in a virtual database of all modified peptides. Exhaustive search leads to a large combinatorial problem, even for a small set of modifications types. Problem (Yates et al. , 1995). Extend the virtual database approach to a large set of modifications. 56

Exhaustive Search for Modified Peptides • YFDSTDYNMAK Oxidation? • • For each peptide, generate all modifications. Score each modification. Phosphorylation? • 25=32 possibilities, with 2 types of modifications! 57

Modified Peptide Identification Problem Goal: Find a modified peptide from the database with maximal match between an experimental and theoretical spectrum. Input: – S: experimental spectrum – database of peptides – Δ: set of possible ion types – m: parent mass – Parameter k (# of mutations/modifications) Output: – A peptide of mass m that is at most k mutations/modifications apart from a database peptide and whose theoretical spectrum matches the experimental S spectrum the best 58

Peptide Identification Problem: Challenge Very similar peptides may have very different spectra! Goal: Define a notion of spectral similarity that correlates well with the sequence similarity. If peptides are a few mutations/modifications apart, the spectral similarity between their spectra should be high. 59

Spectrum Alignment • See 8. 14 and 8. 15 in the text book for one algorithm • Complicated for real spectrums 60

Quality Measure of Mass Spectrometer • Sensitivity • Mass accuracy • Resolution • Dynamic range 61

Ion Types • Some masses correspond to fragment ions, others are just random noise • Knowing ion types Δ={δ 1, δ 2, …, δk} lets us distinguish fragment ions from noise • We can learn ion types δi and their probabilities qi by analyzing a large test sample of annotated spectra. 62

Database Search: Sequence Analysis vs. MS/MS Analysis Sequence analysis: similar peptides (that a few mutations apart) have similar sequences MS/MS analysis: similar peptides (that a few mutations apart) have dissimilar spectra 65

Deficiency of the Shared Peaks Count Shared peaks count (SPC): intuitive measure of spectral similarity. Problem: SPC diminishes very quickly as the number of mutations increases. Only a small portion of correlations between the spectra of mutated peptides is captured by SPC. 66

Ions and Probabilities • Tandem mass spectrometry is characterized by a set of ion types {δ 1, δ 2, . . , δk} and their probabilities {q 1, . . . , qk} • δi-ions of a partial peptide are produced independently with probabilities qi 67

De Novo vs. Database Search: • The database of all peptides is huge ≈ O(20 n). • The database of all known peptides is much smaller ≈ O(108). • • • However, de novo algorithms can be much faster, even though their search space is much larger! A database search scans all peptides in the database of all known peptides search space to find best one. De novo eliminates the need to scan database of all peptides by modeling the problem as a graph search. 68

Probabilistic Model • For a position t δj Ti the probability p(t, P, S) that peptide P produces a peak at position t. • Similarly, for t R, the probability that P produces a random noise peak at t is: 69

Probabilistic Score • For a peptide P with n amino acids, the score for the whole peptides is expressed by the following ratio test: 70

Peak Score • For a position t that represents ion type dj : qj, if peak is generated at t p(P, st) = 1 -qj , otherwise 71

Peak Score (cont. ) • For a position t that is not associated with an ion type: q. R , if peak is generated at t p. R(P, st) = 1 -q. R , otherwise • q. R = the probability of a noisy peak that does not correspond to any ion type 72