1. Protein Identification by Peptide Mass Fingerprinting
Aim:
Introduce concepts behind PMF and some tools that use the technique

2. Protein identification by mass spectrometry - background
Endoproteases cut at specific positions in the amino acid sequence. Trypsin cuts after R (arginine) and K (lysine).
Digestion of a protein with trypsin yields a peptide fingerprint
A mass spectrometer measures mass to charge ratio (m/z) of the peptide masses.

3. Theoretical digest of a protein
K <EK> D
K <ALK> S
K <GWK> I
K <MEK> G
R <GLVK> S
R <YVR> A
K <SPIK> V
R <ADTR> E
K <LEHK> D
K <AMGYR> V
K <GQIVGR> Y
K <EELFR> S
<SIPETQK> G
R <YVVDTSK>
K <DIVGAVLK> A
K <IGDYAGIK> W
K <DIPVPKPK> A
R <VLGIDGGEGK> E
R <EALDFFAR> G
K <ANELLINVK> Y
K <GVIFYESHGK> L
K <CCSDVFNQVVK> S
K <SISIVGSYVGNR> A
R <SIGGEVFIDFTK> E
R <ANGTTVLVGMPAGAK> C
K <VVGLSTLPEIYEK> M
K <LPLVGGHEGAGVVVGMGENVK> G
K <ATDGGAHGVINVSVSEAAIEASTR> Y
K <YSGVCHTDLHAWHGDWPLPVK> L
275.3
330.4
389.4
406.5
415.5
436.5
443.5
461.4
525.6
596.7
628.7
692.7
801.8
810.9
813.9
835.9
893.1
944.1
968.1
1013.2
1136.2
1241.4
1251.4
1312.4
2019.3
2312.4
2418.7
>gi|532319|pir|TVFV2E|TVFV2E envelope protein SIPETQKGVIFYESHGKLEHKDIPVPKPKANELLINVKYSGVCHTDLHAWHGDWPLPVKLPLVGGHEGAGVVVGMGENVKGWKIGDYAGIKWLNGSCMACEYCELGNESNCPHADLSGYTHDGSFQQYATADAVQAAHIPQGTDLAQVAPILCAGITVYKALKSANLMAGHWVAISGAAGGLGSLAVQYAKAMGYRVLGIDGGEGKEELFRSIGGEVFIDFTKEKDIVGAVLKATDGGAHGVINVSVSEAAIEASTRYVRANGTTVLVGMPAGAKCCSDVFNQVVKSISIVGSYVGNRADTREALDFFARGLVKSPIKVVGLSTLPEIYEKMEKGQIVGRYVVDTS
A specific enzyme, here trypsin, cuts the protein into peptides. Trypsin cuts after arginine (R) and lysine (K) and the masses of the peptides can then be calculated. Each protein in a database can be cut in silico to generate a list of peptide masses.

4. MALDI-TOF mass spectrum
A mass spectrum is acquired from a protein digest

5. Peptide mass fingerprinting - principle
A MALDI-TOF spectrum of a protein digest is matched against theoretical digestions of a protein sequence database
A scoring system is used to identify the best database hit.
The idea was first presented at a conference in 1989, and the first papers were published in 1993 by five independent groups.

6. Peptide mass fingerprinting = PMF MS database matching
Protein(s)
Peptides
Peaklist
Mass spectra
Enzymatic
digestion
Result: ranked list of protein candidates
Match
In-silico
digestion
…MAIILAGGHSVRFGPKAFAEVNGETFYSRVITLESTNMFNEIIISTNAQLATQFKYPNVVIDDENHNDKGPLAGIYTIMKQHPEEELFFVVSVDTPMITGKAVSTLYQFLV …
- MAIILAGGHSVR
- FGPK
- AFAEVNGETFYSR
- VITLESTNMFNEIIISTNAQLATQFK
- YPNVVIDDENHNDK …
Sequence
database entry
Theoretical
proteolytic peptides
Theoretical
peaklist

7. Some critical parameters for PMF
Spectrum peak extraction and deisotoping. In the ideal case, all peptide peaks in a spectrum are used for matching, and no other peaks.
Is the protein in the database?
Scoring algorithm
Protein modifications considered
Calibration / mass accuracy

8. From raw data to peak list
Spectrum (m) = baseline (m) + signal (m) + noise (m)
Source: Markus Müller

9. Peptide mass fingerprinting
What you have:
Set of peptide mass values
Information about the protein: molecular weight, pI, species.
Information about the experimental conditions: mass spectrometer precision, calibration used, possibility of missed-cleavages, possible modifications
Biological characteristics: post-translational modifications, fragments
What to do:
Match between this information and a protein sequence database
What you get:
a list of probable identified proteins

10. Data about the protein 2D Gels  molecular weight and pI Importance:
reduce the search space
favours the “good proteins”
Problems:
2D not always available
imprecise measure
proteins may be fragmented
proteins may be modified
Species  databases not always complete, search on close species

11. Data about experimental conditions
Accepted mass tolerance
 due to imprecise measures and calibration problems
Source: Introduction to proteomics: tools for the new biology:. Daniel C. Liebler. Human Press. 2002

12. Need for calibration
Example: Helicobacter spectrum search in Uniprot from ExPASy.
0.1 Da mass tolerance Da mass tolerance Da tol, recalibrated

13. Data about experimental conditions
Number of allowed missed-cleavages
 digestion not always perfect
Chemical modifications  depends on the sample preparation
carboxymethylation of cysteines
oxidation of methionines
etherification of side chain carboxylic groups of glutamic (D) and aspartic (E) acids together with the carboxyl-terminal group

14. Biological characteristics
Post-translational modifications  identification and characterisation of proteins. Use of database annotations
Protein fragments  many forms for each protein, experimental problems
Protein variants  mutations
Depends on the databases available in the identification tools and on the tools too

15. Scoring systems
Essential for the identification! Gives a confidence value to each matched protein
Three types of scores
Shared peaks count: simply counts the number of matched mass values (peaks)
Probabilistic scores: confidence value depends on probabilistic models or statistic knowledge used during the match (obtained from the databases)
Statistic-learning: knowledge extraction from the influence of different properties used to match the proteins (obtained from the databases)

16. Filtering Removal of contaminant peaks before database searching.
- trypsin autolytic peaks
- keratins (hair,skin)
- matrix peaks
- machine artefacts

17. Taking advantage of contaminants!
Identification of contaminants
< 20 ppm
Outlier
Multipoint calibration on contaminants
Source: Karin Hjerno

18. Some PMF tools Non exhaustive list! Tool Source website Aldente
Mascot
MS-Fit
prospector.ucsf.edu/
ProFound
prowl.rockefeller.edu/profound_bin/WebProFound.exe
PepMAPPER
wolf.bms.umist.ac.uk/mapper/
PeptideSearch
PepFrag
prowl.rockefeller.edu/prowl/pepfragch.html
Non exhaustive list!

19. MS-Fit and MOWSE*
NCBInr and other databases, index of masses (many enzymes).
Considers chemical and biological modifications.
Statistic score which considers the mass frequencies.
Calculates the frequency of peptide masses in all protein masses for the whole database. The frequencies are then normalised.
The protein score is the inverse of the sum of the normalised frequencies of matched masses.
The pFactor reduces the weight of masses with missed-cleavages in the frequency computation.
Why talk about MS-Fit before Mascot? Because it’s score algorithm (Mowse) is used by Mascot. Better to explain then in the beginning.
*MOlecular Weight SEarch

20. Mowse score considers the peptide frequencies.

21. Mascot Choice of several databases.
Considers multiple chemical modifications.
0 to 9 missed-cleavages.
Score based on a combination of probabilistic and statistic approaches (is based on Mowse score).
Does not consider SwissProt annotations

22. Mascot - principles Probability-based scoring
Computes the probability P that a match is random
Significance threshold p< 0.05 (accepting that the probability of the observed event occurring by chance is less than 5%)
The significance of that result depends on the size of the database being searched.
Mascot shades in green the insignificant hits
Score: -10Log10(P)

23. Mascot
Input

24. Decoy
Output
Hints about the significance of the score

25. Output Sequence coverage Peptides matched Error function
The first time a peptide match to a query (one spectrum) appears in the report, it is shown in bold face. Whenever the top ranking peptide match appears, it is shown in red. This means that protein hits with peptide matches that are both bold and red are the most likely assignments. These hits represent the highest scoring protein that contains one or more top ranking peptide matches.
Error
function

26. Aldente
SwissProt/TrEMBL db, indexed masses (trypsine and many others).
Considers chemical modifications and user specified modifications.
Considers biological modifications (annotations SWISS-PROT).
0 or 1 missed-cleavages.
Use of robust alignment method (Hough transform):
Determines deviation function of spectrometer
Resolves ambiguities
Less sensitive to noise

27. Aldente – matching and alignment principle
All the couples (peak / peptide)
Experimental masses / peaks
Theoretical masses / peptides

28. The user defines the space area to search in
Dalton error space
Experimental masses / peaks
Theoretical masses / peptides

29. The user defines the space area to search in
Ppm error space
Experimental masses / peaks
Theoretical masses / peptides

30. The user defines the space area to search in
Dalton and ppm error space
Experimental masses / peaks
Theoretical masses / peptides

31. Aldente locates and solves ambiguities
1 peak / 3 peptides
2 peaks / 2 peptides
Experimental masses / peaks
Theoretical masses / peptides

32. Aldente finds the best way to fit the peptides with the peaks
Experimental masses / peaks
Theoretical masses / peptides

33. Aldente – summary
Spectrometer
calibration error
internal error
Experimental masses / peaks
The Hough Transform estimates from the experimental data the deviation function of the mass spectrometer (the calibration error function).
The program optimizes the set of best matches, excluding noise and outliers, to find the best alignment.
Summary
Theoretical masses / peptides

34. Aldente - Input

35. Output
Hints about the significance of the score

36. Information from Swiss-Prot annotation
Information from Swiss-Prot annotation. Processed protein (signal peptide is cleaved).

37. What is the expected information in an identification result?
A summary of the search parameters
A list of potentially identified proteins (AC numbers) with scores and other evidences
Possibilities to validate/invalidate the provided results (info on the data processing, on the statistics, links to external resources, etc.)
Possibilities to export the (validated) data in various formats

38. Hints to know when the identification is correct
The higher the number of masses that match the protein, the better
Good sequence coverage: the larger the sub-sequences and the higher the sequence coverage value, the better
Scores: the higher, the better. Distance to 2nd hit (but there can be more than one protein)
Filter on the correct species if you know it (reduces the search space, time, and errors)
Better when the errors are more or less constant among all peptides found.
If you have time, try many tools and compare the results

39. Protein characterization with PMF data
- Exact primary structure
- Splicing variants
- Sequence conflicts
- PTMs
1 protein entry
does not represent
1 unique molecule
Characterization tools at ExPASy using peptide mass fingerprinting data
Prediction tools
PTMs and AA substitutions
Oligosaccharide structures
Unspecific cleavages
FindMod
GlycoMod
FindPept

40. Detection of PTMs in MS Unmodified tryptic masses  m/z => PTM
624.3
769.8
893.4
994.5
994.5
1056.1
1326.7
1501.9
1759.8
1923.4
1923.4
2100.6
1759.8
Unmodified tryptic masses
624.3
1056.1
769.8
1326.7
1501.9
893.4
2100.6
600
2200
600
2200
 m/z
=> PTM
994.5
994.5
1923.4
1923.4
Tryptic masses of a modified protein
1759.8
1759.8
624.3
624.3
1056.1
1070.1
769.8
769.8
1326.7
1326.7
1501.9
1501.9
893.4
2100.6
2100.6
893.4
600
2200
600
2200

41. FindMod http://www.expasy.org/tools/findmod/ AA modifications DB entry
experimental
options
experimental masses

42. } } FindMod Output unmodified peptides, modified peptides
known in SWISS-PROT
and chemically modified
peptides }
putatively modified
peptides predicted
by mass differences
+ putative AA substitutions