Confident Phosphorylation Site Localization Using the Mascot Delta Score

Large scale phosphorylation analysis is more and more getting into focus of proteomic research. Although it is now possible to identify thousands of phosphorylated peptides in a biological system, confident site localization remains challenging. Here we validate the Mascot Delta Score (MD-score) as a simple method that achieves similar sensitivity and specificity for phosphosite localization as the published Ascore, which is mainly used in conjunction with Sequest. The MD-score was evaluated using liquid chromatography-tandem MS data of 180 individually synthesized phosphopeptides with precisely known phosphorylation sites. We tested the MD-score for a wide range of commonly available fragmentation methods and found it to be applicable throughout with high statistical significance. However, the different fragmentation techniques differ strongly in their ability to localize phosphorylation sites. At 1% false localization rate, the highest number of correctly assigned phosphopeptides was achieved by higher energy collision induced dissociation in combination with an Orbitrap mass analyzer followed very closely by low resolution ion trap spectra obtained after electron transfer dissociation. Both these methods are significantly better than low resolution spectra acquired after collision induced dissociation and multi stage activation. Score thresholds determined from simple calibration functions for each fragmentation method were stable over replicate analyses of the phosphopeptide set. The MD-score outperforms the Ascore for tyrosine phosphorylated peptides and we further show that the ability to call sites correctly increases with increasing distance of two candidate sites within a peptide sequence. The MD-score does not require complex computational steps which makes it attractive in terms of practical utility. We provide all mass spectra and the synthetic peptides to the community so that the development of present and future localization software can be benchmarked and any laboratory can determine MD-scores and localization probabilities for their individual analytical set up. Large scale phosphorylation analysis is more and more getting into focus of proteomic research. Although it is now possible to identify thousands of phosphorylated peptides in a biological system, confident site localization remains challenging. Here we validate the Mascot Delta Score (MD-score) as a simple method that achieves similar sensitivity and specificity for phosphosite localization as the published Ascore, which is mainly used in conjunction with Sequest. The MD-score was evaluated using liquid chromatography-tandem MS data of 180 individually synthesized phosphopeptides with precisely known phosphorylation sites. We tested the MD-score for a wide range of commonly available fragmentation methods and found it to be applicable throughout with high statistical significance. However, the different fragmentation techniques differ strongly in their ability to localize phosphorylation sites. At 1% false localization rate, the highest number of correctly assigned phosphopeptides was achieved by higher energy collision induced dissociation in combination with an Orbitrap mass analyzer followed very closely by low resolution ion trap spectra obtained after electron transfer dissociation. Both these methods are significantly better than low resolution spectra acquired after collision induced dissociation and multi stage activation. Score thresholds determined from simple calibration functions for each fragmentation method were stable over replicate analyses of the phosphopeptide set. The MD-score outperforms the Ascore for tyrosine phosphorylated peptides and we further show that the ability to call sites correctly increases with increasing distance of two candidate sites within a peptide sequence. The MD-score does not require complex computational steps which makes it attractive in terms of practical utility. We provide all mass spectra and the synthetic peptides to the community so that the development of present and future localization software can be benchmarked and any laboratory can determine MD-scores and localization probabilities for their individual analytical set up. Post translational modifications (PTMs) 1The abbreviations used are:CIDcollision induced dissociationETDelectron transfer dissociationETDSAETD with supplemental activationFLRfalse localization rateHCDhigher energy collision induced dissociationLC-MS/MSliquid chromatography tandem mass spectrometryMALDImatrix-assisted laser desorption ionisationMgfMascot generic formatMSAmultistage activationPTMspost translational modificationsQTOFquadrupole time of flight. 1The abbreviations used are:CIDcollision induced dissociationETDelectron transfer dissociationETDSAETD with supplemental activationFLRfalse localization rateHCDhigher energy collision induced dissociationLC-MS/MSliquid chromatography tandem mass spectrometryMALDImatrix-assisted laser desorption ionisationMgfMascot generic formatMSAmultistage activationPTMspost translational modificationsQTOFquadrupole time of flight. of proteins are being actively pursued owing to their broad biological significance. In particular, recent advances in liquid chromatography and mass spectrometry have made the large scale enrichment, identification and quantification of phosphopeptides feasible (1.Beausoleil S.A. Jedrychowski M. Schwartz D. Elias J.E. Villén J. Li J. Cohn M.A. Cantley L.C. Gygi S.P. Large-scale characterization of HeLa cell nuclear phosphoproteins.Proc. Natl. Acad. Sci. U.S.A. 2004; 101: 12130-12135Crossref M. in and phosphorylation in the of phosphopeptides from using and 2004; cell by electron transfer dissociation tandem mass Natl. Acad. Sci. U.S.A. from a for the of from phosphorylated J. mass spectrometry of tyrosine phosphorylation sites in the At the it not to identification and phosphorylation site by of tandem mass spectra of peptide and sites of tyrosine phosphorylation from of and the of phosphorylation to the in a peptide an M. in and phosphorylation in M. site localization of modifications from mass S.A. Villén J. S.A. J. Gygi S.P. for phosphorylation analysis and site to the of the in the tandem MS spectra of phosphopeptides are not to M. phosphopeptide a of site localization for peptide the of is for a fragmentation does not these the mass site localization be significantly The is further by the of sites of in a peptide and the that phosphopeptides are by spectra of the peptide identification as Mascot and identification by using mass spectrometry D. to tandem mass spectra of peptides to in the not for site In ion are for each to peptide which site within a peptide the ion does not to call a phosphorylation site correctly S.A. Villén J. S.A. J. Gygi S.P. for phosphorylation analysis and site collision induced dissociation electron transfer dissociation with supplemental false localization higher energy collision induced dissociation liquid chromatography tandem mass spectrometry laser desorption Mascot generic translational modifications time of flight. collision induced dissociation electron transfer dissociation with supplemental false localization higher energy collision induced dissociation liquid chromatography tandem mass spectrometry laser desorption Mascot generic translational modifications time of flight. of and in phosphorylation site computational have published over the M. in and phosphorylation in M. site localization of modifications from mass S.A. Villén J. S.A. J. Gygi S.P. for phosphorylation analysis and site J. of phosphopeptide from tandem mass M. for and M.A. an phosphorylation site for phosphorylation site analysis of individual for D. a method for phosphorylation site using The two known are the Ascore S.A. Villén J. S.A. J. Gygi S.P. for phosphorylation analysis and site and the similar M. in and phosphorylation in which the fragmentation of phosphopeptides and for the and of in tandem mass The Ascore for the localization for possible site present in a Although the Ascore and to are the Ascore in available is with the used The was phosphorylation data from cell the phosphopeptides with known phosphorylation sites. Both published and in available are for spectra acquired low resolution ion trap and their fragmentation not The method M. site localization of modifications from mass is an of the Ascore for electron dissociation and electron transfer dissociation data in available and In of the is a for that are more applicable to mass spectrometry and fragmentation in available of for phosphorylation site localization be attractive for ion trap collision induced dissociation data and the S.A. Villén J. S.A. J. Gygi S.P. for phosphorylation analysis and site evaluated the of a Mascot that the the two Mascot ion of phosphorylation sites in the peptide by the ion of the The found the MD-score to be in to the The laboratory used the MD-score not validate fragmentation and analysis by mass of site in and spectra of phosphopeptides of In we the ability of the MD-score to the of phosphorylation site localization for commonly used peptide fragmentation of mass using 180 synthetic phosphopeptides with precisely known phosphorylation sites. We found the MD-score to be applicable throughout and provide MD-score thresholds and functions for each fragmentation that as to determine the false localization of phosphorylation site made by In we provide the liquid chromatography-tandem MS data so that of localization software can the of these a data set. we the phosphopeptide available to the community so that any laboratory can determine and MD-score for their analytical set up. a of phosphopeptides M. D. M. J. M. J. of of 180 peptides were synthesized individually by a scale of a peptide the were obtained from peptides were by laser MS and used for further tandem mass spectra of all synthesized phosphopeptides are in were individually by as all peptides were that phosphorylation site were present in any The synthesized peptides in and and and which is similar to the data in 1% J. M. M. site and and of are as and and of the peptides site two which is similar to phosphorylation using as the and The higher of peptides in set with that found in was by the to a number of these peptides for statistical phosphorylated peptides are in in the for and phosphorylated peptides are MD-score thresholds be in these and synthesized phosphopeptides were each to a using a and a to a was within using a from to in and an of The was to the analytical spectra were for followed by spectra for the for was used as the collision and collision energy was and software was used that the data from tandem mass spectra and these into Mascot generic further was were using Mascot with as and and phosphorylation of and tyrosine as was as the and to were The mass of the ion was set to and that of was set to The data was an of the with a J.E. Gygi S.P. for in by mass a of and a of and the of and and was by a to an Orbitrap mass using a trap followed by a analytical was within using a from to in The was to the analytical The mass was in data MS and MS spectra were acquired in the Orbitrap a resolution of after to a of In the were for fragmentation by collision induced dissociation multi stage electron transfer dissociation with supplemental dissociation and of of and peptides were after to a of and a time of The were in the ion with and supplemental peptides were after to a of and a time of was used as the and the time in the ion trap was the for for and for The were in the trap ion MS spectra were acquired a resolution of were after to a of and using a collision energy of were a resolution of in the software was used that the data from tandem mass spectra and these into Mascot generic further was were using Mascot with as and and phosphorylation of and tyrosine as modifications of for not was as the and to were The mass of the ion was set to and that of was set to and for of The data was an of the with a J.E. Gygi S.P. for in by mass a of and a of and the of and and were with and of tandem mass were with and of tandem mass spectra J. M. high peptide identification mass and and of spectra was as M. a for in were in an ion trap mass to a were a made trap followed by a made and was from to in within The mass was using the in data MS and MS spectra were acquired a mass range of after to a of within the the was determined by a followed by fragmentation of and with of and collision software was used that the data from tandem mass spectra and these into Mascot generic further was were using Mascot with as and and phosphorylation of and tyrosine as modifications of for not was as the and to were The mass of the ion was set to and that of was set to The data was an of the with a J.E. Gygi S.P. for in by mass a of and a of and the of and and were with and of tandem mass was as J. M. high peptide identification mass and The MD-score was from Mascot by the the and Mascot ion for phosphorylation site an peptide sequence. The MD-score was by the MD-score by the Mascot ion S.A. Villén J. S.A. J. Gygi S.P. for phosphorylation analysis and site of the Ascore was in the of S.A. Villén J. S.A. J. Gygi S.P. for phosphorylation analysis and site localization for all was by the number of site by the number of site as a of the the combination of the Ascore and the we a to the data in The of the was We Ascore by to the we the highest for each phosphorylation site and MS data mass as as and a for the data are available from the data the Mascot Delta synthetic peptides used in are available from The peptide identification of Mascot are not in a for the localization of a phosphorylation site within a peptide sequence. We the Mascot Delta Score which the of Mascot ion the highest and highest ion for candidate phosphorylation sites an peptide in a be a for site a set of phosphopeptides M. D. M. J. M. J. of of a of 180 phosphopeptides and with precisely known phosphorylation sites and were synthesized individually and by a The of these peptides are similar to found in phosphorylation and be a set of for the of the MD-score for localization The and tandem MS data for each peptide as be the in a analysis of a proteomic In spectra were to the 180 different phosphopeptides peptides to spectra peptide of spectra all Mascot ion Ascore, and MD-score were obtained and the number of as as the false localization were determined the known phosphorylation sites of the synthetic that the of all localization strongly and phosphorylation site the Mascot is not a confident for the of phosphorylation site of the data in that the Ascore a of spectra to the site in a of The MD-score was more and made the of a higher of of all tandem mass the MD-score was that and site localization be the can thresholds that be used for the analysis of similar the Ascore, a of is to 1% which the Ascore made The of the MD-score is which are The 1% for the MD-score is which site were We that the for peptide are very similar to the determined for the Ascore and MD-score S.A. Villén J. S.A. J. Gygi S.P. for phosphorylation analysis and site data that the MD-score and Ascore data and better than the MD-score the fragmentation of and phosphopeptides can be we the Ascore and MD-score be in their ability to with phosphorylation site localization to the different the and data was from the Ascore for peptides 1% of At 1% the Ascore higher sensitivity with the MD-score 1% of At have that the MD-score outperforms the Ascore for by a of 1% of 1% of At the MD-score outperforms the Ascore by a of the show that the Ascore for peptides is strongly the MD-score does not show a the different is than the Ascore for a of MD-score and Ascore for the peptide to the determined that the as the The can be to the of two for of the which of the of site localization for any is a to data to that the MD-score thresholds from the are to obtained from peptide to with the that the from the are very to determined from of MD-score data different fragmentation techniques and for the localization of phosphorylation sites 1% false localization spectra a peptide are used for and the is used for and is from the MD-score spectra a peptide are used for and the is used for and is from the MD-score in a MD-score and Ascore have individual and in using from tandem MS spectra to the site of to the different being that for of the Ascore from the phosphorylated and the are ion are of In Mascot mainly the ion the the phosphorylated the the of the two show a not very to to the two and that the combination of to higher sensitivity 1% higher be the of the two is data to more as we the of two phosphorylation sites within a peptide the of The synthesized phosphopeptide that to for in the MD-score and Ascore for Although can phosphorylation that a more localization in that the two sites are more than At 1% an MD-score of is for phosphorylation sites an MD-score of the phosphorylation sites are further The is for the Ascore and the thresholds are for sites and for sites with that to the peptide in any data set be used with site of individual of peptides of data development of for the of site we the MD-score and using spectra a We the of the MD-score for fragmentation techniques commonly used in In particular, we phosphopeptide the MD-scores of these an and and an mass and The of these are in and 1% and the highest number of phosphopeptides with the phosphorylation localization closely followed by low resolution spectra with supplemental and At the and by were significantly more than the and and the Orbitrap However, methods significantly than and that the high ion mass by and the of in spectra provide more site localization than low resolution and spectra The of all peptides by and is very high However, the of MD-scores the two techniques was the two techniques different have different with to for to be the of the Although spectra from and MD-scores of and the spectra MD-scores of and for the and the that more to better spectra with a the MD-scores for data are for and for the of and for Mascot in and localization in The of all peptides by and is very high and their MD-scores better than for and that the used for site localization are not phosphopeptide are not tandem MS spectra were for each peptide of fragmentation Although to spectra In to the we the data analysis using the The in that mainly to a in that the of thresholds determined be for data with available spectra to the of the MD-score thresholds be to the data in two determine the thresholds for set and the be found for the set. of the number of phosphopeptides available to we to by replicate the MD-score of two acquired and and using a in the of for analysis the in the two that thresholds determined for of the data can be to the We that the of the MD-score using data is in by ion mass we the very data with low high ion At 1% the number of from to peptides site for data an ion trap to obtained from a that are significantly in the we not an in the MD-scores of peptides the two the in localization are owing to The better ion mass than an ion trap and that are in ion owing to their to low mass In the for peptides are than ion the ion trap spectra more than spectra the of the is that data to the of more for site localization and better localization that of tandem mass spectra can have an the of localization that tandem MS spectra to low the peptide identification of proteins from low resolution ion trap spectra by of mass spectrometry 2004; M. peptide identification in by two of mass Natl. Acad. Sci. U.S.A. 2004; 101: is used in the Ascore and and and data that the of phosphorylation localization by the MD-score by and data to site localization is to and Both these the Mascot ion the number of the to and the number of from the over two ion Both not the of the Mascot ion the of the set of 180 synthesized phosphopeptides The data and in that the MD-score can the of these spectra of the peptide from the in which of the was synthesized to all spectra are all a are and site localization the of the the localization the the of the the MD-score in all is which the phosphorylation site in each with is the MD-score to an of the phosphorylation site by in conjunction with phosphopeptide can very be by liquid chromatography using site by the MD-score be possible in an However, peptides to the site identification be possible for the In we have the of a Mascot (MD-score) for ability to localize phosphorylation sites in of using the Mascot ion the MD-score the in Mascot ion the two phosphorylation site by the We a number of and individually synthesized phosphopeptides with precisely phosphorylation sites and similar to found in set of to the of the MD-score in and to the for different a false localization for phosphorylation site made by the Mascot can be for phosphopeptide spectra by commonly used tandem mass spectrometry which we is a to the available set of for phosphorylation site We that the MD-score is not a that it more than In the Ascore S.A. Villén J. S.A. J. Gygi S.P. for phosphorylation analysis and site evaluated the of a Mascot ion is the in the for the two peptides and that by the for low resolution ion trap spectra found it to be to the We the methods to the analysis of phosphopeptide data and in evaluated the of a is the in the for the two The of a the methods are in data the obtained by that the MD-score is significantly than the We a very similar to localization in in the However, that the MD-score significantly outperforms the MD-score and is very similar in to the The for the of the MD-score is that it makes high and low two sites with Mascot of and the MD-score as two sites with Mascot of and the ability to call a correctly by The laboratory used the ion of Mascot for phosphorylation sites from and data fragmentation and analysis by mass of site in and spectra of phosphopeptides of not the statistical made by the Mascot ion are applicable for localization for these two fragmentation the MD-score and Ascore show similar are in The MD-score outperforms the Ascore for tyrosine phosphorylated peptides the Ascore does so for phosphorylation not be that the Ascore was a data set by the phosphorylation sites with high MD-scores not have high and the MD-score and Ascore to a in sensitivity and specificity over using are published the of phosphorylation site identification by using different S.A. Villén J. S.A. J. Gygi S.P. for phosphorylation analysis and site M. for and localization M. in and phosphorylation in M. site localization of modifications from mass S.A. Villén J. S.A. J. Gygi S.P. for phosphorylation analysis and site M. for and M.A. an phosphorylation site for phosphorylation site analysis of individual for D. a method for phosphorylation site using was the of to these methods to the MD-score it can be that and site localization which with which are used for site phosphopeptides in the in proteins with tandem mass 2004; for phosphorylation can be used to phosphorylation site localization of the MD-score is that and statistical are of the of the The MD-scores are in obtained the of the the with modifications in to phosphorylation and MD-score determined for a tandem MS method be used for large data from is attractive it does not require and show that the MD-score can be used for fragmentation However, it be that the MD-score is not an of phosphorylation site localization the MD-score and thresholds are different for fragmentation an not by set of synthetic phosphopeptides with precisely known sites to the MD-score so that false localization can be for any of the fragmentation techniques the by an MD-score of is for site localization in low resolution spectra an MD-score of for low resolution spectra to the of We that any site localization in be to the fragmentation the to for each for phosphorylation site is of the ability to identify the peptide in the of the fragmentation of not all fragmentation techniques and mass are We used the phosphopeptide to the phosphorylation site localization of the MD-score for all commonly used fragmentation The that low resolution spectra are very in correctly the site of phosphorylation cell by electron transfer dissociation tandem mass Natl. Acad. Sci. U.S.A. proteomic of phosphopeptides using electron transfer dissociation tandem mass Natl. Acad. Sci. U.S.A. data have to the number of phosphopeptide identification and localization J. J. J. an for mass J. of and for phosphopeptide J. S.A. Gygi S.P. of the of in phosphorylation and data with these have that than for the identification of phosphopeptides of the time to spectra with spectra J. S.A. Gygi S.P. of the of in phosphorylation so that and be more However, are that the of and for phosphopeptide the of than identification and the data that spectra At the we that is more for an than for an which is in with the was that fragmentation can be as as fragmentation which is commonly to as the method of for phosphorylation site D. J. for the can be to the does not from the low mass to ion trap spectra and of the peptide that are not available from ion trap mass In the ion mass by the Orbitrap analyzer the of the low in spectra Mascot to low for localization more than possible in ion trap the high resolution of spectra and of of which to higher Mascot The was obtained for with supplemental which can be by the more fragmentation of as as the of and which more than At the peptides than with phosphorylation site is the of the synthetic peptides were as that their identification by over for We to the and of data 1% is very for the of site Although the are a of the and the data the of and it have to with ion in the the sensitivity of is very so that we were to in the of and Orbitrap be that the are for site localization and not more identification in the of and more time than However, a very recent that is in a very method for phosphorylation identification J. of large scale with J. Although we the for and phosphorylated the low number of these peptides in data set makes it to of be more for site localization of phosphorylated a data and recent of site in and spectra of phosphopeptides of in J. show that the phosphorylation site of peptides M. in peptide is not a for the of the MD-score phosphorylation site are that be it be that the we present in for the MD-score and the Ascore are in the that are of as possible phosphorylation of that that are further are more to be correctly that a In simple sites the is be the localization for individual peptides sites. be very to as a more to be to a for all However, a very large number of peptides with precisely known sites be to which was the of for to the localization of to peptides sites. Mascot a of of a a peptide in to the time a phosphorylated the the number of sites to phosphorylated the is possible for phosphorylated peptides the is sites and for and a the is sites. We data as as the data in the Ascore and M. in and phosphorylation in S.A. Villén J. S.A. J. Gygi S.P. for phosphorylation analysis and site in and found that is not a in phosphopeptide that these The is for the phosphopeptide in the Ascore the phosphopeptide in the are the we can from data that be of the by Mascot are to have similar it does not a for in and the MD-score in The does more for proteomic modifications are present a large to MD-score thresholds determined can be to data We that the thresholds are very replicate However, as for any localization the that to as data the of tandem mass spectra that thresholds to different is we not data available to the community provide the peptide so that individual can determine MD-score thresholds for their individual analytical data that the MD-score is a for the of phosphorylation site made by which further the of and The of individual synthetic phosphopeptides with precisely known phosphorylation sites as the Ascore, and similar that were large phosphopeptide data in which the sites were not known a is for in which it is practical to validate each phosphorylation site by of the tandem mass in be that of tandem mass spectra be more than an as the The MD-score is applicable to fragmentation techniques and can be obtained from Mascot that Mascot is of the used identification software in the MD-score to their phosphorylation data the for using identification We are all data as as the phosphopeptide available to the community so that any laboratory be to similar of analysis as we and the to their analytical