Modification Site Localization Scoring: Strategies and Performance

Using enrichment strategies many research groups are routinely producing large data sets of post-translationally modified peptides for proteomic analysis using tandem mass spectrometry. Although search engines are relatively effective at identifying these peptides with a defined measure of reliability, their localization of site/s of modification is often arbitrary and unreliable. The field continues to be in need of a widely accepted metric for false localization rate that accurately describes the certainty of site localization in published data sets and allows for consistent measurement of differences in performance of emerging scoring algorithms. In this article are discussed the main strategies currently used by software for modification site localization and ways of assessing the performance of these different tools. Methods for representing ambiguity are reviewed and a discussion of how the approaches transfer to different data types and modifications is presented. Using enrichment strategies many research groups are routinely producing large data sets of post-translationally modified peptides for proteomic analysis using tandem mass spectrometry. Although search engines are relatively effective at identifying these peptides with a defined measure of reliability, their localization of site/s of modification is often arbitrary and unreliable. The field continues to be in need of a widely accepted metric for false localization rate that accurately describes the certainty of site localization in published data sets and allows for consistent measurement of differences in performance of emerging scoring algorithms. In this article are discussed the main strategies currently used by software for modification site localization and ways of assessing the performance of these different tools. Methods for representing ambiguity are reviewed and a discussion of how the approaches transfer to different data types and modifications is presented. Cells respond to elements in their extracellular milieu via a variety of signaling mechanisms, which may be generated in response to environmental cues, electrical stimuli, or chemical messengers. Mostly the signals are propagated by alteration of pre-existing proteins through the addition of post-translational modifications (PTMs) 1The abbreviations used are:PTMpost-translational modificationFDRfalse discovery rateFLRfalse localization rateABRFassociation of biomolecular resource facilitiesiPRGproteome informatics research groupSLIP scoresite localization in peptide scoreCIDcollision induced dissociationECDelectron capture dissociationETDelectron transfer dissociation. 1The abbreviations used are:PTMpost-translational modificationFDRfalse discovery rateFLRfalse localization rateABRFassociation of biomolecular resource facilitiesiPRGproteome informatics research groupSLIP scoresite localization in peptide scoreCIDcollision induced dissociationECDelectron capture dissociationETDelectron transfer dissociation. to key residues to change their structure and activity. These signals are then relayed and amplified until they ultimately manifest in changes in protein expression. It is increasingly clear that mis-regulation of PTMs is a major basis of disease, whether it causes aberrant signaling in cancer, or cytotoxic protein aggregation in neurodegenerative diseases. For this reason, the study of protein post-translational modifications is vital to understanding biological regulation (1Pawson T. Scott J.D. Protein phosphorylation in signaling–50 years and counting.Trends Biochem. Sci. 2005; 30: 286-290Abstract Full Text Full Text PDF PubMed Scopus (498) Google Scholar). post-translational modification false discovery rate false localization rate association of biomolecular resource facilities proteome informatics research group site localization in peptide score collision induced dissociation electron capture dissociation electron transfer dissociation. post-translational modification false discovery rate false localization rate association of biomolecular resource facilities proteome informatics research group site localization in peptide score collision induced dissociation electron capture dissociation electron transfer dissociation. Mass spectrometry-based proteomics has revolutionized the characterization of protein PTMs, in that it has created the first unbiased strategies to identify which proteins are being modified, with what types of modifications, and on which specific residue/s. Of the many types of PTMs used, protein phosphorylation has justifiably received the greatest attention (2Boersema P.J. Mohammed S. Heck A.J. Phosphopeptide fragmentation and analysis by mass spectrometry.J. Mass Spectrom. 2009; 44: 861-878Crossref PubMed Scopus (289) Google Scholar, 3Grimsrud P.A. Swaney D.L. Wenger C.D. Beauchene N.A. Coon J.J. Phosphoproteomics for the masses.ACS Chem. Biol. 2010; 5: 105-119Crossref PubMed Scopus (141) Google Scholar). However, there are many other regulatory modifications such as serine and threonine O-GlcNAcylation (4Hart G.W. Housley M.P. Slawson C. Cycling of O-linked beta-N-acetylglucosamine on nucleocytoplasmic proteins.Nature. 2007; 446: 1017-1022Crossref PubMed Scopus (1077) Google Scholar), lysine or arginine methylation (5Rathert P. Dhayalan A. Ma H. Jeltsch A. Specificity of protein lysine methyltransferases and methods for detection of lysine methylation of non-histone proteins.Mol. Biosyst. 2008; 4: 1186-1190Crossref PubMed Scopus (29) Google Scholar), acetylation of lysine side-chains (6Zhao S. Xu W. Jiang W. Yu W. Lin Y. Zhang T. Yao J. Zhou L. Zeng Y. Li H. Li Y. Shi J. An W. Hancock S.M. He F. Qin L. Chin J. Yang P. Chen X. Lei Q. Xiong Y. Guan K.L. Regulation of cellular metabolism by protein lysine acetylation.Science. 2010; 327: 1000-1004Crossref PubMed Scopus (1458) Google Scholar), or lysine ubiquitination (7Kirkpatrick D.S. Denison C. Gygi S.P. Weighing in on ubiquitin: the expanding role of mass-spectrometry-based proteomics.Nat. Cell Biol. 2005; 7: 750-757Crossref PubMed Scopus (187) Google Scholar) that are all transient and important to study and understand. There are many other PTMs used by the cell, both transient and stable, that affect protein activity. Indeed, different PTMs do not operate independently of each other, so by studying only a single modification type it is impossible to deconvolute their contributions to signaling mechanisms and cellular responses (8Strahl B.D. Allis C.D. The language of covalent histone modifications.Nature. 2000; 403: 41-45Crossref PubMed Scopus (6584) Google Scholar, 9Zeidan Q. Hart G.W. The intersections between O-GlcNAcylation and phosphorylation: implications for multiple signaling pathways.J. Cell Sci. 2010; 123: 13-22Crossref PubMed Scopus (240) Google Scholar). The large-scale analyses of all of these modifications follow a similar strategy, in which an enrichment step (antibody affinity, metal affinity, lectin affinity) is followed by tandem mass spectrometric analysis of the resulting mixture (10Zhao Y. Jensen O.N. Modification-specific proteomics: strategies for characterization of post-translational modifications using enrichment techniques.Proteomics. 2009; 9: 4632-4641Crossref PubMed Scopus (277) Google Scholar). Thousands of modified peptides may be reliably identified in these studies, but the challenge of determining modification site localizations among these peptides is less well addressed. An modification site localization by in the of or of type that are fragmentation between in the peptide that the such an is not or the be the modification be to only in the peptide the modification there be ambiguity as to the of in which a site is and in which there is to between The of site localization and it is for but it be by what is a and what is whether that are of a modification site a that to a a modified or an the is be and which residues in a peptide be as a for a methylation whether that to the mass For peptide of are routinely for The first of these is a measure of for peptide and is a measure of how a of by the this or the the The a false discovery rate is at the data and the of being The is by a in which an is that the of to the and be the Gygi S.P. search for in large-scale protein by mass 2007; 4: PubMed Scopus Google Scholar). a false localization rate be for modification site localization approaches to this and are discussed in a The only in which a be is the modification site localizations are The in which this is the is through the of peptides S. T. phosphorylation site localization using the Cell Full Text Full Text PDF Scopus Google Scholar), but that has to site localizations in a of data in which there only threonine or peptide site localization scoring a search Cell Full Text Full Text PDF Scopus Google Scholar). Although all search engines a score that the certainty of peptide only a currently an score to measure the of modification site localization site localization scoring a search Cell Full Text Full Text PDF Scopus Google Scholar, J. A. a peptide search the PubMed Scopus Google Scholar, J. Zhou H. for Cell 2008; 7: Full Text Full Text PDF PubMed Scopus Google Scholar, Scholar). In the a variety of different to modification site localization S. T. phosphorylation site localization using the Cell Full Text Full Text PDF Scopus Google Scholar, J. J. Gygi S.P. for protein phosphorylation analysis and site PubMed Scopus Google Scholar, F. C. P. in and phosphorylation in signaling Full Text Full Text PDF PubMed Scopus Google Scholar, T. J.D. an phosphorylation site for 2008; 7: PubMed Scopus Google Scholar, S.M. D.L. site localization of modifications mass 2009; PubMed Scopus Google Scholar, D.L. Wenger C.D. Coon J.J. by electron transfer dissociation tandem mass Sci. 2009; PubMed Scopus Google Scholar, T. T. P. C. A. C. and phosphorylation site localization using PubMed Scopus Google Scholar, X. An for peptide tandem mass 2009; 5: Scopus Google Scholar). to the of site localization scoring the only for assessing the site localization the at each in and a on whether they the site localization is a on the of the their to as many as a Indeed, to to site localization S. protein the of Cell 5: Full Text Full Text PDF PubMed Scopus Google Scholar) there many published in which this not they as by the search with The of this the of site localizations that of the data not be in but these to their and it is not these which site localizations on these less is discussed in in a of this the proteome informatics research group of the of a study in on identifying and phosphorylation the to modification site localization the of of software and a or tools. P. W. W. J. Scholar) the of of These main strategies for scoring site localization to the of a that allows site to at or a search score between peptide with different site the are J. J. Gygi S.P. for protein phosphorylation analysis and site PubMed Scopus Google Scholar), F. C. P. in and phosphorylation in signaling Full Text Full Text PDF PubMed Scopus Google Scholar), the in J. Zhou H. for Cell 2008; 7: Full Text Full Text PDF PubMed Scopus Google Scholar), S.M. D.L. site localization of modifications mass 2009; PubMed Scopus Google Scholar), J. Wenger C.D. S. Swaney D.L. Coon J.J. and of and PubMed Scopus Google Scholar), T. T. P. C. A. C. and phosphorylation site localization using PubMed Scopus Google Scholar), of the S. T. phosphorylation site localization using the Cell Full Text Full Text PDF Scopus Google Scholar), the score in Protein site localization scoring a search Cell Full Text Full Text PDF Scopus Google Scholar), and the modification localization score in Scholar) site localization in using an to the strategy, in which site localization are for each post-translational modification on a peptide by the of peptide and by the with the peptide to how their site of of these and the of this article these different site localization their performance and their for data types other phosphorylation data in which the main data type so using these of of the modification site localization scoring or Mass for Mass scoring not score for all score for mass scoring not Protein in each of all score with the of and all score in a An often step in both peptide and modification site localization is which in a to for the there be a mixture of that are the peptide of and other that are by chemical or electrical the are of the of which to is on the basis of an in which only this are an similar to that by the search has T. J.D. an phosphorylation site for 2008; 7: PubMed Scopus Google Scholar). methods are by for The is to a the However, not all the of and are or and that are by fragmentation the of the peptide are in that they are the mass of a of For these a that a to in peptide and site in Protein the in to a for with the of in each of the the in each to a is to the then the each for is the by and the but to being used in of the is in which generated using the and strategies for a of the In each there are a similar of in the However, there are to each that to and of the peptide In this each that be used to the phosphorylation to the threonine it is a different in each the and in the The Protein has with the on the data site localization scoring a search Cell Full Text Full Text PDF Scopus Google Scholar). The differences in to a strategy, but the to and site at a a of site localizations for in which using the other strategies the site localizations The first site localization scoring and similar approaches to score In both the for assessing phosphorylation site in as which as this type of data is with an of is not an step to then the that mass is to be at then are the of of these is in their they both then the The differences between and are that this to for the peptide as a with each site then these for each site localization by the of all of the a of the peptide is modified and to each site in the peptide on this F. C. P. in and phosphorylation in signaling Full Text Full Text PDF PubMed Scopus Google Scholar). In their score only on and that be the for all site localizations are of site localization for all it only a score for the site which is a score between this site localization and the J. J. Gygi S.P. for protein phosphorylation analysis and site PubMed Scopus Google Scholar). In the of the using a score of which to a of or a site being with certainty J. J. Gygi S.P. for protein phosphorylation analysis and site PubMed Scopus Google Scholar). the often a M.P. T. J. W. Gygi S.P. of protein phosphorylation and 2010; Full Text Full Text PDF PubMed Scopus Google Scholar). a of the it is clear that a score the localization to at site determining in the of the of localization The of the by these methods is on the used to data the of the is that all are to be at a of elements and only by a group each other, so are in to be For a at mass be a by the or the other a at be by However, as in each they are or to other site localization this may this of these as of they be as to a by the to certainty In the scoring in to the of the mass of the by the of with is the of is the mass and is the mass of the T. T. P. C. A. C. and phosphorylation site localization using PubMed Scopus Google Scholar). only this for different of an to different of with different for but allows for the of mass for data generated on search is to all of the site localizations in a peptide peptide there is in the search that a measure of site localization by the in for peptide with different site is the that is by S. T. phosphorylation site localization using the Cell Full Text Full Text PDF Scopus Google Scholar), scoring in Protein site localization scoring a search Cell Full Text Full Text PDF Scopus Google Scholar), and modification localization scoring in Scholar). The first both that are the in or be for both on a a in of of score of score In the score is not but on the of their type and the of The score for site localization in score between the localizations to at or between that has a of the of the and are the There are differences between these tools. and the of in and in the with the of and the of used by changes to is the used by as the search this for in peptide scoring is the as is modification localization scoring in the score is by software using the as In addition to the of the in Protein and it that site localization is for all the search for other search software the in to a for analysis that site localization not be routinely search engines need to be to be to site localization for all the search for each However, it is the that the peptide with a different site localization not score in the a of are that are only with site In this site localization score be and so that the for the site localization are to for all site localizations to be In the a data analysis study on the of proteomics in identifying and phosphorylation P. W. W. J. Scholar). The with the data generated the of a of and metal Of the the data for the study only and to identify the in the with false discovery rate and the certainty or ambiguity in the of phosphorylation to of the measure of Gygi S.P. search for in large-scale protein by mass 2007; 4: PubMed Scopus Google Scholar) the of the study be on a However, of the of a metric for false localization rate not be specific on a measure of localization to localization for each peptide and a of their scoring mechanisms and The of the site localization then on on that of the the on the site of localization of only in which ambiguity of of and in which at to the as In the of as in and the not to be to a single or it that the to localization to with the of the be to with on localization in the An be in Although is it the need for a the study the site localization as for this peptide the localization for and for The has only a at and an at that be as different types on the localization of the to or of peptide different software is widely by each to the metric of false discovery rate using a this all on a field the of to peptides and peptides be the Gygi S.P. search for in large-scale protein by mass 2007; 4: PubMed Scopus Google Scholar). on the of peptide and protein be in a by of methods and rate for peptide and protein in 2010; PubMed Scopus Google Scholar). for modification site localization a currently be of the is that peptide with site localizations are not they are similar to the so do not an The approaches all modification of residues that the modification site localization scoring a search Cell Full Text Full Text PDF Scopus Google Scholar). The site localization step be of the peptide as the of residues affect the peptide producing false or false on the is site localization scoring a search Cell Full Text Full Text PDF Scopus Google Scholar). For a localization to be the of the residues need to be the as the and the of the residues to the site localization to follow a similar to the a is to the site of the it is to be as being the modification Although the of the residues and is their in the peptides of residues with is in the be by of and or and However, for peptides the be using such a of be for at by the In addition to a similar of in the with serine and threonine and are for localization scoring they are each in the for many are other and they be to to phosphorylation residues similar used as in study site localization scoring a search Cell Full Text Full Text PDF Scopus Google Scholar). both and residues are residues to phosphorylation in a single peptide then this to localization in a data as the of residues is being by to only type at a but multiple this these types of are to the An be to site localization among residues and residues for a peptide and then the Although and may of the false localization of an peptide the of localizations an data may an that not may be to as a measure of differences in performance and the score used with a as or The of an accepted metric is to to and to by each localization and as to the of localizations in to to a to the of a has by and in determining the of the in the of the be whether a be to be of this is the need to the which is not to be the a modification enrichment step to However, with the of mass and Biol. PubMed Scopus Google Scholar), fragmentation data is in these they be used to this with may a research in the of metric is to the In years a that the of modification site localizations in the of on by The of these S. Zhang a resource for the structure and of post-translational modifications in and PubMed Scopus Google Scholar) and H. C. A. Jensen F. a of phosphorylation PubMed Scopus Google Scholar). These currently the to the localization of the then to the by on site in multiple or data modification identified and in emerging research in the of F. J. the modification PubMed Scopus Google Scholar). a all to the identified through a software and to consistent scoring published modification localizations generated by are by and of the by other the be the and data to on the an modification localization for a peptide is to it data is in the by or of it increasingly for to to the data with the and scoring and on the that they the only to the performance of site localization software is through the analysis of a data in which the are a data created by the analysis of many of each peptide to a large data S. T. phosphorylation site localization using the Cell Full Text Full Text PDF Scopus Google Scholar). In this study they score to for analysis of data on a mass that mass In a study scoring of Protein on this data site localization scoring a search Cell Full Text Full Text PDF Scopus Google Scholar). The that the in of a score to a site localization a site localization score for of peptide with with and with However, of the scoring the with using and using It be that these with in a score for each of these so the of be these the of peptides with site localizations be However, these a in the performance of these tools. only and that the modification be on scoring and scoring may other types on the type scoring and scoring to or may be the for score and scoring site localizations for a of in the at in the of the of these not to affect the of the site In a the of and it other localization the to the on the field in data generated by dissociation methods and T. T. P. C. A. C. and phosphorylation site localization using PubMed Scopus Google Scholar). sets generated used to the score to specific score then to data sets of enrichment of a of a to that the other by as as Although this by that the peptide data may be and may not the of in the biological in the of a widely accepted metric for an on each data these to a in a a a scoring it be to in the of of to pre-existing algorithms. are to identify modification are for peptide to the of the data is to measure the with mass The mass allows of it is less that a be as a It allows which false the mass for and is less for this data the of mass the and the is what J. Wenger C.D. S. Swaney D.L. Coon J.J. and of and PubMed Scopus Google Scholar). However, the that all mass be increasingly as the of mass the of the is then the of a score change for data be for the of the of mass is not then a score may the but be all of the of the mass for site The on search scoring is to as search engines may mass in their they of mass in their scoring then an the measure of for all data for the score by a of with to S. T. phosphorylation site localization using the Cell Full Text Full Text PDF Scopus Google Scholar). It that a score be for the mass search that be reliably In that scoring are a similar for and mass data the score the mass data to as a has at mass but the of a score to be relatively to is that with mass data it may be to a for as are less to be an in false The of the that site localization software are data there to be a in performance of these software other PTMs are the are on the of that between then the on whether the modification a scoring to then a different used for a modification that not it is to this as there are currently large data sets of peptides with PTMs other to a data of modification is to a data to peptides that only then for modification of on an modification data modified identified this data is However, only that a single or there only of which with a which is not to to this is the of Although the not a defined the on of the modification site are other and A. of protein O-GlcNAcylation using electron transfer dissociation mass on Sci. 2009; PubMed Scopus Google Scholar) of modified peptides in this data a single this for a data of a single threonine or site localization scoring a search Cell Full Text Full Text PDF Scopus Google Scholar). that the of a modification a to the to modification what types of modifications the greatest The modifications there is the for site localization modifications that are in the mass O-linked modifications such as and these are less of an in and modifications that on many modifications that on both peptide and acetylation on protein and peptide as well as lysine these are in fragmentation of the mass An important to these modification localization is that the which types a so the modification is on a that not as a a for lysine but not for or then the may be unreliable. For it is not to reliably the site of modification and different strategies are to this type of In the of and a measure is for each site and it is to the to what they to In the of scoring all site localizations that are the score as are which of these is In the of or score the localization is with a what the site localization of the Protein is that for with site localizations the on the peptide it the different site localizations the of the and allows of is as of the it for the The of modification site scoring software has the and of being Of the types of strategies for these the search approaches be to of mass and the different fragmentation of different modifications, so be However, the of a score using these strategies has to data to data In all as there is currently to measure a for the has to the that a consistent a data to