A Simulated MS/MS Library for Spectrum-to-spectrum Searching in Large Scale Identification of Proteins

Abstract

Identifying peptides from mass spectrometric fragmentation data (MS/MS spectra) using search strategies that map protein sequences to spectra is computationally expensive. An alternative strategy uses direct spectrum-to-spectrum matching against a reference library of previously observed MS/MS that has the advantage of evaluating matches using fragment ion intensities and other ion types than the simple set normally used. However, this approach is limited by the small sizes of the available peptide MS/MS libraries and the inability to evaluate the rate of false assignments. In this study, we observed good performance of simulated spectra generated by the kinetic model implemented in MassAnalyzer (Zhang, Z. (2004) Prediction of low-energy collision-induced dissociation spectra of peptides. Anal. Chem. 76, 3908–3922; Zhang, Z. (2005) Prediction of low-energy collision-induced dissociation spectra of peptides with three or more charges. Anal. Chem. 77, 6364–6373) as a substitute for the reference libraries used by the spectrum-to-spectrum search programs X!Hunter and BiblioSpec and similar results in comparison with the spectrum-to-sequence program Mascot. We also demonstrate the use of simulated spectra for searching against decoy sequences to estimate false discovery rates. Although we found lower score discrimination with spectrum-to-spectrum searches than with Mascot, particularly for higher charge forms, comparable peptide assignments with low false discovery rate were achieved by examining consensus X!Hunter and Mascot, results by mass and score results comparable to achieved evaluating consensus and Mascot. with data using X!Hunter with the simulated library than and than with the We that simulated libraries the search available for spectrum-to-spectrum searching and that the approach used in consensus strategies for search Identifying peptides from mass spectrometric fragmentation data (MS/MS spectra) using search strategies that map protein sequences to spectra is computationally expensive. An alternative strategy uses direct spectrum-to-spectrum matching against a reference library of previously observed MS/MS that has the advantage of evaluating matches using fragment ion intensities and other ion types than the simple set normally used. However, this approach is limited by the small sizes of the available peptide MS/MS libraries and the inability to evaluate the rate of false assignments. In this study, we observed good performance of simulated spectra generated by the kinetic model implemented in MassAnalyzer (Zhang, Z. (2004) Prediction of low-energy collision-induced dissociation spectra of peptides. Anal. Chem. 76, 3908–3922; Zhang, Z. (2005) Prediction of low-energy collision-induced dissociation spectra of peptides with three or more charges. Anal. Chem. 77, 6364–6373) as a substitute for the reference libraries used by the spectrum-to-spectrum search programs X!Hunter and BiblioSpec and similar results in comparison with the spectrum-to-sequence program Mascot. We also demonstrate the use of simulated spectra for searching against decoy sequences to estimate false discovery rates. Although we found lower score discrimination with spectrum-to-spectrum searches than with Mascot, particularly for higher charge forms, comparable peptide assignments with low false discovery rate were achieved by examining consensus X!Hunter and Mascot, results by mass and score results comparable to achieved evaluating consensus and Mascot. with data using X!Hunter with the simulated library than and than with the We that simulated libraries the search available for spectrum-to-spectrum searching and that the approach used in consensus strategies for search of in is a in available is peptides by by of peptides by and fragmentation in a mass fragmentation spectra to as MS/MS spectra) from peptide sequences with the observed fragment strategy for matching MS/MS to a peptide and searching using mass the in the of the search programs peptide sequences to spectra in of by from observed and matching the sequences against peptides in a protein or peptide sequences from the protein simple spectra a of and fragment generated by peptide and matching the fragment to used to evaluate observed and use of or searching using mass the in the of the An alternative strategy direct spectrum-to-spectrum matching of spectra against reference MS/MS in a library and of mass spectra library search for that search and libraries similar in by from a reference use and to the and using that evaluate the of the to for the observed fragment However, spectrum-to-spectrum matching more use of fragment ion intensities as as other than the and library searching is and to peptide mass libraries for protein is to spectra and sequences the this reference libraries of peptide MS/MS from data with libraries of available from the of and the spectra) of peptide MS/MS spectra from using Chem. and the spectra) peptide mass libraries for protein peptides also a of and peptides. of spectrum-to-spectrum matching using MS/MS libraries from observed spectra is that peptide sequences In the spectra in the mass by charge in generated peptide a of peptide sequences that were the of in using a with and from data of MS/MS Chem. simulated spectra generated for in generated peptide a of peptide sequences that were the of in using a with and from data of MS/MS Chem. simulated spectra generated for the available reference libraries a small of spectra for peptides. from of low found in alternative with low a spectra available in a library to a is to the that peptide is to of spectrum-to-spectrum matching with spectrum-to-sequence strategies of in the sizes of search and the of a protein in the In this study, we that spectrum-to-spectrum searching by using a library of simulated spectra for peptides in the spectra were generated using a kinetic fragment ion intensities for peptide fragmentation in the MassAnalyzer program by Z. Z. Prediction of low-energy collision-induced dissociation spectra of Chem. Z. Prediction of low-energy collision-induced dissociation spectra of peptides with three or more Chem. MassAnalyzer MS/MS generated by kinetic to peptide to a library of previously MS/MS in as fragmentation and and for peptide is also fragment and of the and a simple of charge the is in the spectra of and charge from Z. Prediction of low-energy collision-induced dissociation spectra of Chem. that the simulated spectra good discrimination by to peptide We that the spectra used to evaluate in a program to of MS/MS spectra of peptide MS/MS assignments using a to use fragment ion intensities in evaluating or assignments to simulated spectra discrimination with data and of with low from the search a of assignments that were normally using score of peptide MS/MS assignments using we for using a library of simulated spectra of peptides from for spectrum-to-spectrum simulated library is than the reference a search comparable to that used by spectrum-to-sequence searching of protein were to this library to the of the and and to searches of approach also of or libraries for searches search strategy for in protein by mass to to evaluate the of peptide of simulated spectra for spectrum-to-spectrum searching performance available reference libraries and more searching of data of a mass with and in by Chem. or a mass with data were used in this data and peptide assignments from a data set of MS/MS spectra a protein from using ion mass as previously and in by Chem. used of false discovery rate false simulated library by simulated library for simulated library for and X!Hunter and data MS/MS spectra a of the of of of protein and peptides were with a of in MS/MS were and charge by MS/MS the with a of and a of for with and of for the MS/MS with three and of with of for data MS/MS spectra peptides from protein of by protein data using as with from to in three and mass and search programs used in this were by and peptide mass libraries for protein and by and of peptide MS/MS spectra from using Chem. X!Hunter with using and BiblioSpec a with using were with spectrum-to-sequence search of by a with and and An approach to mass data of peptides with sequences in a protein with using were for the ion using or X!Hunter and for fragment using for and or for fragment using X!Hunter is in the and is by the in the In the fragment ion mass is by fragment ion mass with that by the BiblioSpec also ion mass and used for simulated library searching were and for and and X!Hunter observed ion to is the peptide and mass that for charge we the simulated library by charge and set for charge to results In to mass we to and for protein library is to other MS/MS by a score ion BiblioSpec a score reference library and and X!Hunter the score score and to a in the were or In the X!Hunter reference library from observed is as of the score this available for the simulated a of used for searches with the observed and simulated In the of the data a for the of MS/MS the search results the use of as the the of the a data libraries of observed spectra for X!Hunter and BiblioSpec were from the simulated libraries were the protein protein for or a protein in of sequences generated for peptides with mass and to and in using a with and from data of MS/MS Chem. to were for that a of were this in generated peptide a of peptide sequences that were the of in using a with and from data of MS/MS Chem. of spectra the program MassAnalyzer the simulated spectra as the simulated spectra a for simulated spectra were to the for X!Hunter and for and the for of to for the X!Hunter library and set to the of and sizes of the libraries and the library used in this libraries of peptide sequences used in this library of in peptide sequences limited by mass of to charge to and is by protein the of spectra of of to of is and for in the BiblioSpec in peptide sequences limited by mass of to charge to and in using a with and from data of MS/MS Chem. is by protein the of spectra of is and for in the BiblioSpec in a MS/MS spectra were by with using the for data and the for were as for X!Hunter and Mascot, and a to data to the and and for the of spectra and used by X!Hunter and BiblioSpec and and for the of spectra and were and in by Chem. of a with used in for charge were than of spectra in data used in this library search libraries that is to libraries using searches against the simulated library were by the data set and library as previously for searching in using a with and from data of MS/MS Chem. of data and simulated libraries were by charge and by with to the mass searching with data we for X!Hunter that the searches and a by and charge X!Hunter for the X!Hunter searches the to the of the and by from the X!Hunter simulated library is available for from the X!Hunter with the and results by from the to evaluate the use of the simulated spectra for spectrum-to-spectrum matching were using library search X!Hunter and the and the X!Hunter reference library of observed MS/MS to as were from and implemented as and we a simple to the of and of the search and of and in by Chem. In this a reference library from MS/MS spectra from a data set data peptide spectra were used as the data set for a MS/MS as data to in the reference library generated from the BiblioSpec MS/MS spectra to reference library X!Hunter spectra to reference with three and that the lower of matches by of spectra by data in X!Hunter of the with were were by a to or were the the X!Hunter MS/MS with three and We the to the that were in reference library were for data of the spectra by X!Hunter a the than of the the other to were used. this the that were We the simulated spectra generated by MassAnalyzer substitute for reference library spectra in spectrum-to-spectrum we generated from observed with a a of the spectra by simulated spectra as in spectra were from used to data set data the data set using or the of this to the to peptides the of other spectra the peptide sequences and charge found in the peptide sequences and peptides with or other library of spectra in with simulated the and also the spectra with simulated spectra generated from sequences of the peptides the to the MS/MS spectra in the data spectra the data and of X!Hunter and were in the search to a peptide sequences were to and spectra for and library assignments to or other we to the assignments as and the as from and false we by other search assignments to the library search assignments. We to of the assignments by the data set against the decoy library a of matches to or a for a for this data of the we the assignments were or to a and library searches the were as or of the and low spectra that were more to by of the matches to the search as and as the matches to the library were as and were or assignments were with the assignments to by as from the decoy library of assignments by X!Hunter than that observed by searches against a decoy MS/MS to sequences of the of MS/MS with a the X!Hunter and results to of peptides for the in spectra from the of the spectra in the the protein used by peptides from the of peptides. the of matching protein higher for than for the with the observed that the observed false assignments to in of were lower than that the library mass or charge of the data a more comparison of the performance of the library in this a using in the results were as and false or false and the to a peptide from the score to the of that the for in use of the data set of matches as of the the and library in this is a of MS/MS assignments in the searches that by peptides in search the assignments for the library search were as MS/MS small and that assignments the or that assignments to peptides for the search were is similar to that of the library the in the comparison of the search the results the by the small of assignments to by and the of the simulated is to that the for in the and charge available to in the assignments to the library search reference spectra were from data as by the of low mass in the assignments to the search the simulated spectra the spectra low of the with for comparable of the simulated and spectra the library and spectra for other we that with a library of simulated spectra from the search to a comparable to protein a of matches normally with the simulated spectra libraries were and were to the We used MassAnalyzer to the simulated spectra for peptides in the protein as as a generated from protein peptides with mass and charge or were and peptides to as previously in using a with and from data of MS/MS Chem. library to to the of the search libraries for X!Hunter and BiblioSpec and libraries and library a that for the of the and with library by charge and mass simple of the In to to the library the and from a the search X!Hunter and BiblioSpec libraries is the of fragment we to as and the of fragment in the spectra that in the the from by BiblioSpec the or of the in the and the to the of the of in searches of the data set against the library using BiblioSpec to the or fragment and with fragment the is set to that the is the to the of the from the spectra with in the reference In BiblioSpec searches against the of assignments to with the of in library with in the of assignments from a decoy search discrimination were observed as by the in score for assignments to other assignments and with assignments. with charge for as the of from to were observed with or of searches using the data program and and libraries used by X!Hunter and is the reference library generated from observed and simulated libraries generated by MassAnalyzer for X!Hunter and library is peptides in the the of for in the library and libraries used by X!Hunter and is the reference library generated from observed and simulated libraries generated by MassAnalyzer for X!Hunter and library is peptides in the the of for in the library and libraries used by X!Hunter and is the reference library generated from observed and simulated libraries generated by MassAnalyzer for X!Hunter and library is peptides in the the of for in the library and libraries used by X!Hunter and is the reference library generated from observed and simulated libraries generated by MassAnalyzer for X!Hunter and library is peptides in the the of for in the library and libraries used by X!Hunter and is the reference library generated from observed and simulated libraries generated by MassAnalyzer for X!Hunter and library is peptides in the the of for in the library of of MS/MS of a by the search of MS/MS as to in the peptides of to other the decoy to in the as a decoy of MS/MS as to in the peptides of to other the decoy to in the as a decoy MS/MS were from the data set that peptide assignments and in direct comparison results using and of decoy MS/MS as to in the peptides of to other the decoy to in the as a decoy and libraries used by X!Hunter and is the reference library generated from observed and simulated libraries generated by MassAnalyzer for X!Hunter and library is peptides in the the of for in the library of MS/MS of a by the search MS/MS as to in the peptides of to other the decoy to in the as a decoy MS/MS were from the data set that peptide assignments and in direct comparison results using and in a X!Hunter searches of the data set against with higher of assignments to were In of and assignments charge with BiblioSpec to score by a for in to the the results with this to use X!Hunter in We the of searches against and the data set against spectra of were to peptide sequences peptide sequences and were from we the data set to MS/MS to searches using the of the data set against assignments to and searches against assignments against decoy assignments low of assignments to in the of assignments lower than the observed with the decoy library the were as in the as in the the of assignments achieved with comparable to that achieved in searches of the MS/MS to using the of the simulated library for more from data assignments to were by MS/MS were the peptide by libraries the spectra were from fragmentation of with and In of the were from and by of the searches in of assignments the peptide to evaluate the of the assignments and spectra charge the matches were found in the spectra and peptide sequences In spectra and peptide the with by the of spectra for the peptide sequences or charge in other sequences by searches were in the search against the reference library or and to to for the simulated spectra in in We also that assignments observed spectra matches than simulated In matches were observed in charge spectra) and peptide spectra for and charge were in We to as other spectra in higher than the the simulated spectra and the small of fragment against a search or the and in a and the use of for the reference peptide sequences were by peptide sequences were from and of we we the assignments using the this in assignments as for this we a the of the that score for MS/MS to were comparable searches with the to by the the and assignments In assignments were in a the and and and assignments in discrimination that to and score and of alternative to search results with We and that evaluating consensus search programs and assignments in discrimination is and in by Chem. by results from MS/MS search approach the from In study, we consensus and a false discovery rate were to the false of for data We a similar consensus approach to with spectrum-to-spectrum An for in a consensus approach is good is achieved for sequences by the search programs or we the of assignments achieved with searches against the X!Hunter searches against of X!Hunter and searches of the data set MS/MS to peptide sequences in the and and the sequences were were peptides in with and sequences in that a consensus approach X!Hunter and using the low discrimination consensus and X!Hunter with consensus and in results in this the assignments for the is for the of the of the mass results similar to of Mascot. consensus matches to peptide consensus matches peptide In this a false is as a consensus that is were in searches using the decoy protein and or decoy decoy matches by consensus and by with the of assignments for in the and for and results that consensus X!Hunter and of is lower than the achieved by and of the data set using with and consensus with or the results of the search and decoy for score were used in the assignments. to peptides with than and with charge mass by ion mass to to spectra with charge to and sequences with that the in for the decoy search the search is to the that the as a higher of in the decoy of MS/MS of peptide sequences in of MS/MS of peptide sequences in for the protein assignments and for the from the decoy is for the and searches assignments for MS/MS to the of consensus the mass is for the protein assignments and for the from the decoy is for the and searches assignments for MS/MS to the of consensus the mass is of MS/MS of peptide sequences in of MS/MS of peptide sequences in for the protein assignments and for the from the decoy is for the and searches assignments for MS/MS to the of consensus the mass is for the protein assignments and for the from the decoy is for the and searches assignments for MS/MS to the of consensus the mass is of MS/MS of peptide sequences in of MS/MS of peptide sequences in for the protein assignments and for the from the decoy is for the and searches assignments for MS/MS to the of consensus the mass is for the protein assignments and for the from the decoy is for the and searches assignments for MS/MS to the of consensus the mass is of MS/MS of peptide sequences in of MS/MS of peptide sequences in for the protein assignments and for the from the decoy is for the and searches assignments for MS/MS to the of consensus the mass is for the protein assignments and for the from the decoy is for the and searches assignments for MS/MS to the of consensus the mass is In the results of the search and decoy for score were used in the assignments. to peptides with than and with charge by ion mass to to spectra with charge to and sequences with that the in using a with and from data of MS/MS Chem. in for the decoy search the search is to the that the as a higher of in the decoy of MS/MS of peptide sequences in for the protein assignments and for the from the decoy is for the and searches assignments for MS/MS to the of consensus the mass is in a We the of assignments mass model to estimate the of peptide by MS/MS and Chem. assignments with observed mass to of the mass were this the of MS/MS matches by with a in the of peptide sequences of the MS/MS were assignments the for or a to a the the of decoy or decoy library searches to and by and for consensus by and three by we three assignments for and assignments for We that evaluating consensus X!Hunter and Mascot, with a for mass to a for data with the of peptide We also mass more as a mass the search than in using mass of MS/MS assignments with or consensus as of the in search In decoy or decoy library searches using mass of or similar of false by consensus We this to that consensus the of matches against of sequences of the search strategy or search is more to consensus assignments with a mass of the We the consensus approach with a mass to a data set of a of from by and by results were with searches MS/MS peptide were by using and mass the of consensus MS/MS with of and consensus MS/MS with of of protein by peptide is a of the of the false higher in In the and were by low results of a data set by mass of the the peptide and for search and consensus and using the or decoy and consensus X!Hunter and X!Hunter searches and decoy were MS/MS in this data search the assignments were to charge to and ion mass to MS/MS with the of and and peptide of protein assignments by is and in the of that by peptide is were found in the search of were protein by for search and consensus and using the or decoy and consensus X!Hunter and X!Hunter searches and decoy were MS/MS in this data search the assignments were to charge to and ion mass to MS/MS with the of and and peptide of protein assignments by is and in the of that by peptide is were found in the search of were protein by in a observed for the data set were similar to the of data the false discovery rate lower with than consensus and higher for of assignments by consensus by peptide and by and assignments to consensus by peptide and by by charge of a peptide or peptide sequences to a protein of the using mass of the searches using and to a search of for the of is higher than that of the consensus the search more by peptide that were by by is with that is achieved by evaluating consensus search the advantage of or In spectrum-to-spectrum searching against libraries to consensus results comparable to the spectrum-to-sequence consensus and with An advantage with evaluating the search for of the consensus with the data set and in searches against the by and Mascot, searches against by X!Hunter to of the the comparison of is the protein and spectra the results to the search However, in using a with and from data of MS/MS Chem. that the of a search is to the of the spectrum-to-spectrum searching against the library in search for the advantage in a consensus search that good results with spectrum-to-spectrum searching using a simulated library MS/MS spectra simulated by the kinetic program approach three spectra were generated for peptides from protein the search by the sizes of reference comparable to search of protein used in spectrum-to-sequence peptide in a using the approach for by of from the with X!Hunter searches were than a consensus approach with to assignments. decoy libraries that to protein a for the of assignments in searches and of false discovery for matching in the used by spectrum-to-sequence a of were spectra in the reference libraries of observed MS/MS spectra with the simulated spectra and the library of simulated spectra using the protein used for and strategy the simulated and observed spectra by of and false assignments. However, spectra by peptide that by library a of peptide of searches against and with the data set a in assignments peptides peptide sequences a by using observed the more in In the that assignments were found in the the spectra for sequences were in that simulated spectra in in a comparable of assignments were achieved with searching the protein and with X!Hunter searching that a library of the MS/MS spectra used in spectrum-to-spectrum the other we observed discrimination and assignments of X!Hunter and to a low false discovery rate using score as for the protein is X!Hunter and the set of peptides and program used a we the use of a consensus Although results were achieved with the approach a advantage in the for X!Hunter of this to use a consensus approach for of assignments. with the also that consensus with mass a more alternative to the used a the protein by peptide MS/MS spectra We were that similar is achieved in and X!Hunter the low discrimination for the X!Hunter that good that this approach is of the to the score in comparison with other for of the discrimination to of simulated spectra that discrimination with also that the of the peptides is the that of peptides that and is a use of score as is also in other advantage of a consensus approach is that score An in using the simulated the with the peptide sequences to the search with of a of MS/MS in a data set to a of the protein in using a with and from data of MS/MS Chem. approach in the simulated spectra library and also used to search by simple We and also that by results from search programs and low false discovery achieved with score of a of the MS/MS in a data set and in by Chem. by results from MS/MS search we of the simulated library in a that the spectra used in search results to of search with search and the simulated spectra MS/MS by consensus with consensus in in spectrum-to-spectrum of of and lower the of is the use of for the library and the is and and the of in the peptide the of in the However, to the simulated spectra used in this and to is also that the MassAnalyzer of MS/MS spectra We previously that a of of the simulated spectra is to with of peptide MS/MS assignments using In the of the of higher charge and of of charge in the fragment is that from more in has found that the of the model for MS/MS for of the is that the simulated spectra libraries of observed MS/MS limited of the of peptides. of the search with X!Hunter and the to a library for of results that a in the simulated spectra and this a search for We to Z. for with particularly in to the of as as of the in In we for with X!Hunter of and of for with and of for data