Glcan ID user guide - GlycopeptideID
Contents
1. The tags include fragments N HN S SH and SHN The idea is to get rid of low quality or non glycopeptide spectra which do not contain low mass glycan fragments Filtering by series The spectra which MS peak differences do not match to any of the monosaccharide H N S F masses As previously the idea is to get rid of low quality or non glycopeptide spectra applied i i PRIES rics www appliednumerics com Glycopeptide ID User s Guide Combining MS spectra which match to the same precursor same charge and RT m z differences less than given tolerances are combined MS peaks are pasted to a same peak list and intensity values are not summed so that the resulting spectra contain all the peaks from all the input spectra Marking NeuAc tags Spectra with peaks matching to the NeuAc mass tags are marked The tags include masses of S SH and SHN The idea is that if a spectrum matches to a sialylated glycan it should also have fragments matching to the mass tags Also if the spectrum matches to a non sialylated glycan it should not have matching NeuAc tags Match Peptides Peptide sequences which match best to input spectra are searched The search has two steps 1 Optional filter to limit the possible peptides It is assumed that a MS spectrum contain peaks which can be mapped to a peptide total mass For example an intact peptide connected to a single HexNAc is typically a major peak in the N glycopeptide spectrum 22. GlycopeptideID User s Guide Version 0 9 beta 28 02 2014 Hannu Peltoniemi hannu peltoniemi appliednumerics fi PREC ics Glycopeptide ID User s Guide Contents NC 3 Abbreviatio E 3 T o Yu o dui e E 3 2 WOT le 4 Create Peptide DB amp Fetch Protein iii 5 Create Glycan DB amp Convert GlycomeDB nmesserseaserrsvessenssersersersstensenssosserseesseensensensersstensenssenserseesseensensenseesstensennsenr 5 Pre Process PEA A dd 5 Match Pepe SA RA AA AAA e 6 Match GIy ans at AAA nta 7 Estimate False Discovery Rate ENEE 7 3 np t formats EE 8 MS Spectr E 8 PHOUCUINS PE OPRIMA noo o non 9 GIYCansaasssvasuN ee a deel edt Mie ee den each seedeiee 9 4 GlycopeptidelD Web Va er ere 10 SES HE 11 Tables EXCEL E 11 INTE 13 2 applied ics www appliednumerics com Glycopeptide ID User s Guide Preface The GlycopeptideID software is developed by Applied Numerics in collaboration with the University of Helsinki The aim of this document is to give a short overview about the functionality of the GlycopeptideID and to describe the input and output formats The software is currently under development and it is possible that the actual features may deviate slightly form the ones presented here For questions about the software you may mail to Hannu Peltoniemi hannu peltoniemi appliednumerics fi Abbreviations H Hex Hexose N HexNAc N Acetylhexosamine S NeuAc N Acetylneuraminic acid sialic acid F Fuc Fucose De
3. Quick sweep to rank peptides with the number of matching fragment ions and calculation of score values for the peptides with highest number of matching fragments The sweep is done against pre calculated fragment database and is fast enough to be applied even without the filtering step When calculating peptide scores the intensity is not taken into account as the peptide fragments usually have low intensity compared to glycan and glycopeptide fragments Details of the score calculation are given at the Match Glycans section Peptide match can be guided with several options for example which ion series are searched for required number of fragment matches etc It is also possible to limit the output to include only the peptides which can be combined with glycans to match the precursor total mass Peptides can be modified Variable terminus modifications N and C term can be included with a rather small penalty to the calculation time Also variable amino acid modifications can be included but they can be searched only when the peptide total mass filter is applied and they can increase the calculation time considerably Fixed modifications can be included by modifying the amino acid masses when the peptide database is generated applied i i PR iF erics www appliednumerics com Glycopeptide ID User s Guide Match Glycans Glycans which match to precursor and potential peptide mass differences are searched and scored against MS spectrum The sear
4. agments Estimate False Discovery Rate False discovery rate FDR estimation for the peptide sequences of the highest ranking glycopeptides is implemented with the target decoy approach The idea is to apply identical workflow Fig 1 against target forward and decoy inverted peptide databases Random false matches are assumed to have an equal probability to match target or decoy i e to have highest scoring peptide in target or decoy databases The FDR rate is estimated as a filtered competition between target and decoy ref The precursors are ordered with peptide score and the FDR for matches above a given score value is the percentage of decoy matches relative to target matches When reporting the matches the decoy hits are assumed to be removed filtered from the result applied i i PR iF erics www appliednumerics com Glycopeptide ID User s Guide 3 Input formats The input data contains MS spectra and protein and glycan databases Examples of all the input data types can be found at the GlycopeptidelD Data page MS spectra The supported MS spectra formats are pkl Waters and mgf Matrix Science It is required that the spectra is deconvoluted to a monoisotopic peak lists with resolved charge states The pkl format files can be generated with MaxEnt3 algorithm www waters com and the mgf files with Mascot Distiller www matrixscience com The pkl format file is assumed to contain lines prec m z prec in
5. ch spectrum applied i i PRIES rics www appliednumerics com Glycopeptide ID User s Guide Proteins The protein input is the UniProt original flat text format protein files The GlycopeptidelD site contains pre loaded the entries of the reviewed Swiss Prot human proteins New protein files can be downloaded from UniProtKB site www uniprot org When downloading set the file format to text or flat text A small set of proteins can also be fetched with GlycopeptideID Protein DB web page The proteins are identified by UniProt ID names for ex TRFE HUMAN or by UniProt accessions for ex P02787 Each ID or accession has to be separated by a new line or comma Glycans The glycan structure tables can be generated from the GlycomeDB www glycome db org database dump file or they can be imported by a user The glycan structure table file is a tab delimited table which contains at least three columns structure composition and source The structure column contains glycan structures in modified IUPAC condensed format using single letter glycan names H N S F or G and without linkage for ex SHNH SHNH HNN The composition contains glycan compositions for ex S2H5N4 and source is an optional web link to a database where the glycan is fetched from The glycan fragment databases can be created from the glycan structure lists However the fragment database can also be imported by a user The database is composed by a zipped folder con
6. ch contains steps 1 Glycans which match to MS precursor and peptide mass difference are searched Glycans can be structures or compositions Structures are given as a database and compositions are generated by de novo with a given range of monosaccharide units 2 Glycan fragments free and attached to peptide are matched and scored with MS spectra The theoretical structure fragments for each glycan structure are given in a glycan database The theoretical composition spectra contain fragments that can be selected from the given composition i e the spectrum is composed by all the fragments that any glycan structure with the given composition could produce Scoring the theoretical and measured fragments is done with a statistical score defined by 1 S Cogio Px logso P where Py is the probability that a random set of fragments would have as many or more shared peaks with the measured spectrum as the ranked glycan and P is the probability that by randomly selecting the observed number of shared peaks the same or higher amount of intensity can be covered The total score for a glycopeptide is given by the sum of glycan and peptide scores Note When scoring a composition it is possible that the matches contain fragments that are difficult to interpret originating from a single credible structure Especially in case of suspicious compositions and with compositions with nearly equal scores it is important to validate the matching fr
7. found at the www appliednumerics com site The service is available for registered and anonymous guest users For guests there are some limitations for the size and storing time of user data If there is need for a large scale analysis several thousand MS spectra it is possible to set up a dedicated server for the purpose The site contains web pages Home Home page basic instructions Search Glycopeptide search options and search start Importing new spectra Results Viewing the calculated results Downloading the results Peptide DB Generation of a peptide DB Fetching proteins from UniProt Glycan DB Generation of glycan fragment DB Data Viewing and deleting stored user data Help Documentation The pages are rather self explanatory with some option help and use tips 10 applied i i PR it erics www appliednumerics com Glycopeptide ID User s Guide 5 Result files All the result files can be downloaded in a single zip file The file includes some excel templates an output directory containing all the data files and a Results html file containing links to the main result files Tables Excel The main results are mapped to Excel templates glycopeptides xls and peptides xls The templates contain links to xml and pdf type data which locate at the output folder To open linked xml files with Excel it is required that Excel is set as a default viewer for xml type data select an xm
8. glycans The program is ran as an automated batch process i e for a given inputs and options the output is generated without interactive steps For practical reasons the workflow is divided to three separate tasks so that the peptide and glycan database generations are separated from the rest of the workflow Glycans Protein names GlycomeDB UniProt ID AC CONVERT FETCH GLYCOMEDB PROTEINS MS spectra Glycan Proteins pkl mgf structure list UniProt text PRE PROCESS CREATE CREATE SPECTRA GLYCAN DB PEPTIDE DB Processed Peptide spectra DB MATCH PEPTIDES Matched peptides Decoy peptide DB MATCH PEPTIDES Matched peptides Matched glycopeptides FDR ESTIMATION Glycopeptides with FDR estimate Matched glycopeptides Fig 1 Sketch of the GlycopeptideID workflow The optional tasks are shown in lighter hue applied CS www appliednumerics com Glycopeptide ID User s Guide Create Peptide DB amp Fetch Proteins Peptide database is created from imported UniProt flat text format protein files http www uniprot org The GlycopeptidelD web page contains also a fetch tool suitable for importing a small number of specific proteins directly from the UniProt Proteins are digested with trypsin with a given maximum number of missed cleavage sites Peptides with glycan sites are stored as a peptide database Glycan sites are defined by a given amino acid rule or by the UniProt database annotation For false disco
9. l file with Windows Explorer select Open With gt Choose Default Program and select Excel Peptides peptides xsl contain the matched peptides The possible columns are precursor Running precursor index The order how the spectra has been in the input file mz intensity charge Precursor m z intensity and charge values first_scan last_scan Precursor retention time start and end scan number start_time end_time Precursor retention time start and end time in minutes ms2_spectrum_file Link to spectrum data in xml format If there is a matching molecule the data contains also matching fragments combined_precursors Precursor indexes that have been combined to single file Optional protein Protein name UniProt ID peptide_mass Monoisotopic peptide mass in Da start end Peptide sequence start and end indexes N site N site N glycosylation index N site DB N site index in UniProt DB sequence Peptide sequence dmz Mass charge m z difference between precursor and given molecule ion Charge carrier usually H 11 applied i i PRIS erics www appliednumerics com Glycopeptide ID User s Guide rank peptide Peptide rank for the given precursor score peptide Peptide score hits peptide Number of matching peptide fragments I peptide Relative amount of intensity explained 0 none 1 all by Matching b and y ion numbers also ions a c x z if the
10. oxyhexose G NeuGc N Glycolylneuraminic acid 1 Introduction The purpose of the GlycopeptidelD tool is to identify intact glycopeptides measured with LC MS2 experiments To accomplish that it is required that the glycopeptide spectra contain both peptide and glycan fragments The peptide sequences are identified from the peptide fragments and the glycans from the glycan and glycopeptide fragments The outcome is a set of best matching peptides and glycans for each MS precursor The identified glycan can be a glycan composition or a glycan structure 1 order The tool input is deisotoped MS spectra with identified charge states The peptides are searched against protein database and glycans against glycan database or de novo generated glycan compositions The computational methods resample the ones published by Joenv r et al 2008 The main difference is that in the GlycopeptideID glycan structure database is used and no de novo structures are evaluated Also the peptide identification is developed further and includes for ex other protein modifications Some background to the methods used is also given at Peltoniemi et al 2009 2013 applied i i PRIES rics www appliednumerics com Glycopeptide ID User s Guide 2 Workflow The computational functionality of the GlycopeptideID can be shown as workflow given at the Fig 1 The input is glycopeptide MS spectra which is matched against peptide and glycan databases and or de novo
11. score for given precursor score all Total score defined as a sum of peptide and glycan scores hits all Total number of hits sum of peptide and glycan hits I all Total explained intensity sum of matching peptide and glycan intensities decoy hits Number of decoy hits with a current or higher peptide score value The value is determined only for best ranking rank 1 glycopeptides Optional FDR False discovery rate for peptides of best ranking glycopeptides with a given or higher peptide score Optional filtered_by_decoy Label 0 1 marking if the precursor has a larger score against decoy 1 or target 0 database The FDR is calculated with an assumption that the decoy hits are filtered out from the results Optional References Joenvaara S Ritamo I Peltoniemi H Renkonen R N Glycoproteomics an automated workflow approach Glycobiology 2008 18 4 339 349 Peltoniemi H Joenv r S Renkonen R 2009 De novo glycan structure search with CID MS MS spectra of native N glycopeptides Glycobiology 2009 19 707 714 Peltoniemi H Natunen S Ritamo I Valmu L Rabina J Novel data analysis tool for semiquantitative LC MS MS2 profiling of N glycans Glycoconj J 2013 30 2 159 170 13 applied i i PR iF erics www appliednumerics com
12. taining a file of glycan structures glycan_db txt and one fragment file for each structure All files are tab delimited type tables The glycan_db txt file contains the three columns present in the structure table and a new column ms2 theoretical spectrum file which contains the names of the theoretical fragment files for each structure The fragment file contains columns name structure structure map composition mass type and cuts The name column contains the ion name and the structure is the fragment structure string The structure_map is a string composed by 0 s and 1 s so that it maps the fragment structure to the whole unfragmented structure string 1 monosaccharide present 0 absent The composition is fragment composition and the mass is the fragment monoisotopic mass The type is the fragment type so that fragments with type 1 are attached to a peptide and the fragments with type 2 are not attached to a peptide The cuts value is the maximum number of glycosidic cuts that are required for a given fragment The columns name and structure map are not required and composition and structure columns are used only as strings to annotate peaks The cuts column is used to calculate the maximum possible number of dropped H20 relative to residual masses caused by glycosidic cleavages applied i i PRIES rics www appliednumerics com Glycopeptide ID User s Guide 4 GlycopeptidelD web server Public version of the GlycopeptidelD is
13. tensity prec charge m z intensity m z intensity m z intensity m E The first line is precursor ion m z intensity and charge and the following lines are fragment m z and intensity values The fragments are assumed to be converted to charge 1 peaks one charge carrier mass included for ex one hydrogen One file can contain several spectra separated be one empty line The mgf format is assumed to contain lines BEGIN IONS PEPMASS prec m z prec intensity CHARGE prec charge SCANS scan start scan end RTINSECONDS time start time end m z intensity charge m z intensity charge END IONS Where prec m z prec intensity and prec charge are precursor m z intensity and charge values and the scan start scan end and time start time end are precursor elution time in scans and seconds The precursor intensity value is optional and is set to 100 if missing Also the scan and retention time lines are optional The precursor data may contain other lines starting with a character but are not taken into account Lines starting with a number are assumed to contain MS2 peak data and the first three columns are m z intensity and charge values If the charge is missing the value is set to 1 for each peak If the charge column is present the table may contain also other columns but they have no effect on the calculation Lines starting with BEGIN IONS and END IONS define the beginning and the end of ea
14. very rate FDR estimation a decoy database is created by reversing the database peptide sequences The peptide database generation have a separate web page and is run separately of the rest of workflow The reason is that the calculation can be time consuming and typically there is no need to update the peptide database every time the whole workflow is run Create Glycan DB amp Convert GlycomeDB Glycan database is created from the GlycomeDB www glycome db org database dump xml file which is pre loaded to the GlycopeptideID web server The GlycomeDB file is converted to a structure table with a given taxonomy glycan core and monosaccharide residues Glycan fragment database is generated from the structure table with a given maximum number of glycosidic cleavages ring fragmentation is not included It is possible to import user defined structure tables and also glycan fragment databases Importing new or modified databases may be beneficial if there are some special requirements not taken into account with the current service As with the peptides the glycan database generation have separate web page and is run separately of the rest of workflow Pre Process spectra The pre processing contains several optional filtering combining and tagging tasks Filtering Filtering by given precursor RT retention time charge or m z values Filtering by mass tags The spectra which do not contain any of the typical MS2 mass tags are filtered
15. y are matched spectrum Link to pdf picture of MS spectrum with matching fragments If modifications are searched the output can contain columns modif Modification name variable aminoacid modification modif_site Modification site index modif mass Modification mass value modif_AA Modified aminoacid modif_number Number of modifications N_term_modif N term modification name N_term_modif_mass N term modification mass C_term_modif C term modification name C_term_modif_mass C term modification mass Glycopeptides glycopeptides xls contain the matched glycopeptides In addition to the peptide columns the possible columns are mass add Additional mass added to composition total mass Optional composition Glycan composition structure Glycan structure ms2 theoretical spectrum file Name of the file that contained the theoretical fragments used when matching the structure source Glycan source can be a string or url to a web page ppm Glycopeptide and precursor mass difference in ppm 12 applied i i PRIIEGrics www appliednumerics com Glycopeptide ID User s Guide rank glycan Glycan rank for given precursor score glycan Glycan score hits glycan Number of glycan glycopeptide fragment matches I glycan Relative explained intensity by glycan glycopeptide fragments 0 none 1 all rank all Rank defined by total
Download Pdf Manuals
Related Search