IsobariQ 1.1 User Manual - Norwegian Proteomics Society
Contents
1. For ratios higher than the median What is the probability of obtaining a ratio this high or higher by chance alone And for ratios lower than the median What is the probability of obtaining a ratio this low or lower by chance alone This implies that the statistical test to be applied has to be one sided for each question The probability can be calculated using the Gaussian probability function 16 significance an CO t e 2 at x which returns the upper tail of the distribution Since the test is one sided for each question the values obtained will range from O to 0 5 The significance level is up to the user but IsobariQ calculates the limits shown in the distribution graph at a 0 05 These limits can easily be calculated as follows Lower limity 95 M 1 645 SDiower Upper limitg os M 1 645 SDupper Where M is the median of the log of the normalised ratios These limits are now in log space and need to be transformed before applied to ratios The constant 1 645 is the z score from the normal probability table that ensures 5 probability 2 4 Graphs Many different graphs exist depending on quantification mode selected Original data is shown in blue while normalised data is shown in orange By right clicking in the graph the image can be copied or saved The XY data can also be retrieved for pasting into any spread sheet application and further plotting Here the graphs are explaine2. gt i N 0 500 1000 1500 2000 2500 3000 Rank of Peptide Intensity Plot to see relation between peptide ratio and peptide intensity 23 2 4 15 Peptide Variability vs Peptide Angle Score Peptide Variability vs Peptide Angle Score co n 2582 Peptide Variability gt o o 9 p UJ ul ta N gt N 0 5 10 15 20 25 30 Peptide Angle Score Plot to see if the peptide variability is comparable proportional to our own metric the angle score Graph shows that the two correlate but are not identical 2 4 16 M A M A plot n 29774 1 2 3 4 5 A log10 treated logi0 control x 0 5 Minus vs Add plot Used to look for intensity dependent trends banana shape M is the quant point ratio while A becomes the average intensity for a ratio Actually it is identical to 2 4 9 but kept separate in the list due to its well known name 24 2 5 The Q IPTL Module Pairing K6PP_HUMAN Label verification Sequence lonscore Expect ppm N term Modification AA Modification position mass C term Modification IPTL labelled 4 FLEHLSGAGK 55 23 FLEHLSGAGK 45 73 0 000566677 0 00 _Succinyl2H 4 N term K10 _IMID MDHI_fight K PLEHLSGAGK 95 2 37868 0 00 _Succinyl N term K10 _IMID 2H 4 MDHI_heavy K Mascot result one aa rs telalel Sees filtering K G c s e ec Reset quantitation 26 1854 25 12 24 36 0 5 Ratio Type Massl MassH Intl IntH Matchl Match H 1 W W 2576 VH 3693 3714
3. IsobariQ 1 1 User Manual November 2010 Magnus Arntzen The Biotechnology Centre University of Oslo Christian J Kohler The Biotechnology Centre University of Oslo Copyright 2010 The Biotechnology Centre University of Oslo Permission is granted to copy and distribute this document together with the IsobariQ software Contents L NOGE O oaaaescennacstanqsauuecsanceiaeatpecesesenqs cine eassuet E E E E E 4 1 1 Reporter ions techniques iTRAQ and TMT seessssssssseessssrrereessrreressrrrreessrtreessrreressrerrresererresseeeree 4 1 2 IPTE N A E E E E E soa sesucusneeeesuears 4 2 Working With ISO AO seccina S 4 2 1 Ms talationrandliCEnCE seee vaueteans uetescavesunstuntheneavelaansenumeenevnienereutaeiuncs 4 2 1 1 Installation of R vsn and RSefVe 2 ee nee ee ne eee eee 4 2 2 POO CUD eC iesene E 5 2 2 1 S pported File a6 10 6 cs een ee eee ner n TENN eee er ere E T 5 2 2 2 Ate NOON V E eeina E AS 5 2 2 3 Filtering of Mascot Results and Identification of IPTL Labelling ccceeecccessseceeeeeeees 10 2 3 Quantification Normalisation and Significance sessesesseesesreesrrrssrrresrrerssrrressreresrereserereseerese 10 2 3 1 Calculation of Peptide Ratios and Variability c ccccccccsssececcsseccceeseceeseeseceseeeceeeeeeeeeeas 11 2 3 2 Normalisation of Peptide Ratios sccecsaiaccerducansr dedesSecnparernaoseaaertdocwerdedanecenaaese ven aeexeehecnesieans 13 2 3 3 Calculation of Protein Ratios an
4. 1 08454 2 8 5 R IsobariQ preferences cone x RServe Path to Rserve exe within R bin C Program Files R R 2 11 1 bin Rserve exe For IsobariQ to communicate with R the path to Rserve needs to be set Note This is the path to the Rserve exe that was copied to the R bin folder see section 2 1 1 32 OO 10 11 12 13 14 I5 16 References Ong S E Blagoev B Kratchmarova l Kristensen D B Steen H Pandey A Mann M Stable isotope labeling by amino acids in cell culture SILAC as a simple and accurate approach to expression proteomics Mol Cell Proteomics 2002 1 5 376 86 Thompson A Schafer J Kuhn K Kienle S Schwarz J Schmidt G Neumann T Johnstone R Mohammed A K Hamon C Tandem mass tags a novel quantification strategy for comparative analysis of complex protein mixtures by MS MS Anal Chem 2003 75 8 1895 904 Ross P L Huang Y N Marchese J N Williamson B Parker K Hattan S Khainovski N Pillai S Dey S Daniels S Purkayastha S Juhasz P Martin S Bartlet Jones M He F Jacobson A Pappin D J Multiplexed protein quantitation in Saccharomyces cerevisiae using amine reactive isobaric tagging reagents Mol Cell Proteomics 2004 3 12 1154 69 Koehler C J Strozynski M Kozielski F Treumann A Thiede B Isobaric peptide termini labeling for MS MS based quantitative
5. 3635 1411 y7 2 JH 4002 4043 3830 2112 y4 3 fH 4339 4360 289 3A yss yB Interaction 4 JH 4874 4909 15090 6038 y5 5 H 604 66 12629589T y6 6 eTA 205 2 5755 bis 7 M 7315 7416 11280 5685 y7 8 fA 8295 833 1010 1239 b Ratio gt 9 8665 87S 4835 1978 y g 10 V 3627 UH 995 9840 642 177 9 visualization 1 000 u v WSO H 10123 10168 322 161 b9 5 Highlighting extreems Plotting and anni When IsobariQ is in IPTL mode and a protein has been double clicked the Q IPTL module will open and all the peptides belonging to that protein can be viewed and validated Detailed information like annotated MS MS spectrum possible sequence matches IPTL pairs and quantification is available The user can interact with the quantification events something that causes IsobariQ to recalculate on the fly 25 2 6 The Q Reporter Module Quantitative Peptide Labeling Using Reporter lons Mascot CPSM_MOUSE result Idendficatons Spectrum Information Sequence lonscore Expect ppm N term Modification AA Modification position mass C term Modification filtering YAIAK 32 55 0 00275173 0 00 iTRAQS amp plex N term K6ciTRAQMplex K YLALAK 17 29 0 0923858 0 00 iTRAQ4plex N term KGciTRAQ4 plex K Quanstabon Re quantitate Reset quantitation Type Ratio Natio Int Numerator Int Denominator 1 116 115 0995 1 367 394984 3 396917 3 Ratio 2M 117 115 0549 0599 2181047 396917 3 visualization Interaction MSMS scans Ma
6. proteomics J Proteome Res 2009 8 9 4333 41 Choe L D Ascenzo M Relkin N R Pappin D Ross P Williamson B Guertin S Pribil P Lee K H 8 plex quantitation of changes in cerebrospinal fluid protein expression in subjects undergoing intravenous immunoglobulin treatment for Alzheimer s disease Proteomics 2007 7 20 3651 60 Ow S Y Salim M Noirel J Evans C Rehman Wright P C iTRAQ underestimation in simple and complex mixtures the good the bad and the ugly J Proteome Res 2009 8 11 5347 55 R Development Core Team R A Language and Environment for Statistical Computing http www R project org 2010 Gentleman R C Carey V J Bates D M Bolstad B Dettling M Dudoit S Ellis B Gautier L Ge Y Gentry J Hornik K Hothorn T Huber W lacus S Irizarry R Leisch F Li C Maechler M Rossini A J Sawitzki G Smith C Smyth G Tierney L Yang J Y Zhang J Bioconductor open software development for computational biology and bioinformatics Genome Biol 2004 5 10 R80 Huber W von Heydebreck A Sultmann H Poustka A Vingron M Variance stabilization applied to microarray data calibration and to the quantification of differential expression Bioinformatics 2002 18 Suppl 1 S96 104 Perkins D N Pappin D J Creasy D M Cottrell J S Probability based protein identification by searching sequence databases using m
7. tools IsobariQ is started by double clicking the lsobariQ exe file lsobariQ is licensed under the GNU General Public Licence version 3 and parts of the software mainly libraries are licensed using compatible licences To view all licences and the external libraries click F4 or see the help menu in IsobariQ 2 1 1 Installation of R vsn and Rserve For optimal usage we recommend installing the statistical software R This enables IsobariQ to communicate with R and perform variance stabilizing normalization VSN of the data See section 2 3 2 2 of this manual for details R can be downloaded from www r project org Once installed run R and install Bioconductor by typing source http b1loconductor org biochite R biocLite This will install several necessary packages and amongst them vsn If for some reason vsn is not installed it can be installed with the command biocLite vsn In addition to R and vsn you have to download and install a frontend server called Rserve For Windows download the stand alone windows binary from www rforge net Rserve files and unpack this into a directory of choice The executable Rserve exe needs to be copied to the bin folder of R usually located in C Program Files R due to its dependence on the file R dll For IlsobariQ to perform VSN the path to Rserve needs to be set in preferences lsobariQ was developed and tested with R version 2 11 1 and Rserve version 0
8. 6 2 2 2 Importing Data 2 2 1 Supported File Formats lsobariQ supports data generated by Mascot www matrixscience com Mascot is a proteomics search engine that uses mass spectrometry data to identify proteins from primary sequence databases and access to the Mascot dat files is required Mascot stores its search results in a text based file located on the mascot server usually in the folder C INETPUB MASCOT data yyyymmdd We recommend copying these dat files locally before importing to IsobariQ for performance reasons Note Only Mascot dat files originating from MS MS searches are supported Peptide mass fingerprints or error tolerant searches cannot be loaded in IsobariQ Note If the Mascot dat files are large gt 200 MB importing them in IsobariQ may be time consuming To improve loading time one can apply filters See Data Import Wizard below for details IsobariQ stores the search results in memory while loading and if gt 2 GB of RAM is used it may close unexpectedly 32bit version of IsobariQ IsobariQ was designed to use memory for storage for performance reasons 2 2 2 Data Import Wizard Importing data into IsobariQ is done by selecting Import Data from the File menu This will open the Data Import Wizard Note Not all settings applied in the wizard can be changed after data loading These settings will be visible in preferences but disabled To change one of these parameters the data has to be
9. RAQ4PlexCorrectionMatrix txt See M Vaudel ef al Proteomics 2010 20 650 670 for more details Back Next gt Cancel Help The measured intensities of the reporters have to be corrected for isotopic overlap interference see details below section 2 3 1 2 This can be achieved in IsobariQ using correction matrices as previously described by Vaudel and colleagues The matrix can be stored in a text file and imported into IsobariQ during loading Some example matrices for iTRAQ and TMT can be found in the conf folder but have to be adjusted for every different kit 2 2 3 Filtering of Mascot Results and Identification of IPTL Labelling Mascot usually suggests 10 sequences hits for each MS MS spectrum and ranks them according to their ion score Which of the sequences Mascot uses depends on the protein context If a hit falls below the ion score cut off set by the user while loading data the hit is not loaded If the removed hit was the one used by Mascot for this protein the peptide will not show as a part of the protein and the protein score is recalculated If Mascot suggests a sequence that is labelled according to the selected IPTL labels for example N terminally succinylated light and the C terminal lysine being labelled with MDHI heavy then this hit is marked as IPTL labelled When all hits have been checked for IPTL labelling IPTL pairs will be created if they have matching sequences but opposite IPTL labels succinyl
10. aph to see if the peptide CV is related to the peptide intensity The peptide intensity is not the precursor intensity but the sum of fragment intensities that are assigned to the sequence 2 4 6 Peptide CV vs Rank of Peptide Intensity Peptide CV vs Rank of Peptide Intensity 200 n 2582 150 7 100 50 0 500 1 000 1 500 2 000 2 500 3 000 Rank of Peptide Intensity 19 Often it is useful to plot versus the rank of intensities instead of the intensity itself From this graph one can easily conclude that the CV seems homogeneous for all peptide intensities i e it is not linked to the peptide intensity It is a bit more difficult to get to that conclusion in the previous graph 2 4 7 Quant Point CV vs Mean Intensity Quantpoint CV vs Mean Intensity 140 n 29774 120 100 80 gt O 60 40 20 0 1 2 3 Logi0 Mean Intensity Quant point CV is something completely different from peptide CV The quant point is one ratio calculated from two intensities The quant point CV is here the CV calculated from these two intensities The mean intensity is the mean of these two intensities So in this plot the CV or variance between two intensities are larger for low intense data than for high intense 2 4 8 Quant Point CV vs Rank of Mean Intensity Quantpoint CV vs Rank of Mean Intensity 140 N 29774 120 0 5000 10000 15000 20000 25000 30000 Rank of Mean Intensity When the same see previou
11. ass spectrometry data Electrophoresis 1999 20 18 3551 67 Vaudel M Sickmann A Martens L Peptide and protein quantification a map of the minefield Proteomics 2010 10 4 650 70 Karp N A Huber W Sadowski P G Charles P D Hester S V Lilley K S Addressing accuracy and precision issues in iTRAQ quantitation Mol Cell Proteomics Golub G H Van Loan C F Matrix computations 3rd ed Johns Hopkins University Press Baltimore 1996 p xxvii 694 p Galassi M GNU scientific library reference manual 2nd ed Network Theory Bristol 2002 p xvi 601 p Cox J Mann M MaxQuant enables high peptide identification rates individualized p p b range mass accuracies and proteome wide protein quantification Nat Biotechnol 2008 26 12 1367 72 Nesvizhskii A Aebersold R Interpretation of shotgun proteomic data the protein inference problem Mol Cell Proteomics 2005 4 10 1419 40 33
12. ated heavy and MDHI light This will only occur however if this is the sequence Mascot uses for this peptide Special IPTL Feature If in a peptide Mascot cannot identify the reverse sequence i e only identified in one direction in Mascot lsobariQ will try to force find the opposite sequence If this sequence is found in the MS MS spectrum and produces a number of valid ratios to be calculated this value can be set in the import wizard this hit is stored as force found and an IPTL pair is generated Only MS MS spectra with IPTL pair identifications can be quantified For reporter ion quantification all MS MS spectra with detected reporters can be quantified Whether a peptide can be quantified or not is annotated in the table by a green tick mark or a red X respectively Peptides that are quantifiable due to the above described force found method are annotated with an orange tick mark 2 3 Quantification Normalisation and Significance The workflow of computations in IsobariQ is as follows i Quantify each peptide individually For IPTL this means quantifying each fragment pair individually and calculating a peptide median and peptide variability For reporter ions this means correcting the measured intensities with the correction matrix and calculating the different ratios between the reporters All at the MS MS level ii Normalise the peptide ratios Normalisation corrects for unequal loading and experimen
13. ccinyl 2H 4 N term Lysine light label _IMID MDHI_light K Lysine heavy label _IMID 2H 4 MBHI_heavy K C terminal light label lt none gt C terminal heavy label lt none gt Standard IPTL refers to the original published method which includes cross wise peptide labelling of N terminals and lysines This technique requires a Lys C digest causing relatively long peptides where ETD fragmentation might be superior to CID Tryptic IPTL is a novel derivative of this method utilizing trypsin as the protease and specific C terminal labelling 0 Import Data Wizard Reporter Ions Applicable for 7TRAQ TMT and custom reporter ion techniques Presets Number Of Reporter Ions 4 Name Monoisotopic Mass 114 114 111228 115 115 108263 116 116 111618 11 117 114973 Using the reporter ions technique the user can select from a set of preset methods or choose the number of reporters manually Up to eight reporters are supported 0 Import Data Wizard Reporter ions Applicable for 7TRAQ TMT and custom reporter ion techniques Number Of Ratios B Ratio Numerator Ratio Denominator ii4 114 hid lt tc The user can manually select the number of ratios for IsobariQ to report maximum 7 O Import Data Wizard Reporter ions Applicable for TRAQ TMT and custom reporter ion techniques Use Reporter Ion Correction Matrix Filename C TsobariQ conf fiT
14. culated as follows 2 1 Mceorrected C x Mmeasured The inverse of C is calculated using first LU decomposition Gaussian elimination with partial pivoting followed by solving the system Ax b for each column of the identity matrix All computations are performed using the GNU Scientific library For example using the above correction matrix and the following 114 117 reporters with intensity 1000 each the calculations become 1 08 0 02 0 00 0 00 1000 1055 M _ 0 07 1 09 0 04 0 00 X 1000 _ 984 EE 0 00 0 07 1 09 0 05 1000 976 0 00 0 00 0 05 1 08 1000 1034 We can now see that the uncorrected ratio 115 114 1 but after correcting for impurities and isotopic overlap we get the ratio 0 9 2 3 2 Normalisation of Peptide Ratios Normalisation is performed by using the complete distribution of calculated ratios That is for IPTL all quant point ratios since one peptide can have several ratios and for reporter ions it is the peptide ratios In IsobariQ different settings in preferences determine whether a peptide ratio should contribute to the protein ratio or not For example peptide being razor or unique or Mascot rank and bold normal typeface See section 2 8 1 for more details Only the ratios that are selected to contribute to the protein ratio are used for calculating the normalisation 13 In IsobariQ the user can choose between No Normalisation Division by Median Variance Stabilising Normalisa
15. d Variability cccccccccsssececeesececeesscceeeeseceseeeceeeeeneeeeas 15 2 3 4 Calculation of Protein Ratio Significance and Limits for Up and Down Regulation 15 2 4 O ee ee re Re RO a eee 17 2 4 1 Protein Ratio Distribution HiStORram ccccccsssccccesseccccseccceesccecaesececseeceesegeccessusecetsuneess 17 2 4 2 Peptide Ratio Distrib tion Histogram ccs jncaccsiscndvisssvdeconanccasovaitosesnsetesubesctdesanavassnnnedeanvenas ens 18 2 4 3 Quant Point Ratio Distribution Histogram esssssssessseeresrrressrrrssrrerererresreeresereesereresereeserere 18 2 4 4 Quant Point WN So accesses a E a 18 2 4 5 Peptide CV vs Peptide Intensity ceccccssecccssecceseccesceceeececeeceeeneeeeencessesceseaecesseceneness 19 2 4 6 Peptide CV vs Rank of Peptide Intensity sseseesessensssrerssrrreesrrresrrererrrresererssreresrrereseerese 19 2 4 7 Quant Point CV vs Mean Intensity sssssenssssenssrssessesseserssrsrrsrrsrrssroressesresseesesseeseeseeseesee 20 2 4 8 Quant Point CV vs Rank Of Mean Intensity sseesenseseenesreesssrerserrerererressreresrrresereresereesseere 20 2 4 9 Quant Point Ratio vs Mean Intensity sssssenssnsenssnsenerrssrsressrssrssrssessessessessessessessessessessee 21 2 4 10 Quant Point Ratio vs Rank of Mean Intensity esssesoseeusesrenssrenssrrererrrresrrerssrrresrrererreresene 21 2 4 11 Peptide CV vs Peptide Angle SCOPE cccccccsscccccsseccccensccceeesececceec
16. d briefly 2 4 1 Protein Ratio Distribution Histogram Protein Ratio Distribution Histogram n 383 Frequency 17 This graph shows the histogram of protein ratios 2 4 2 Peptide Ratio Distribution Histogram Peptide Ratio Distribution Histogram 400 n 2582 Frequency N UJ Qo O amp en 1 10 100 Ratio This graph shows the histogram of peptide ratios It should be similar to the one for protein ratios but usually contains more data points and thus is smoother 2 4 3 Quant Point Ratio Distribution Histogram Quant Point Distribution Histogram 4 000 n 29774 3 000 2 000 Frequency 1 000 0 01 0 1 1 10 100 Ratio This graph shows the histogram of quant point ratios IPTL only It should be similar to the one for protein and peptide ratios but usually contain even more data points and thus is smoother 2 4 4 Quant Point Intensities 18 Quantpoint Intensities n 29774 UI log10 Treated Intensity N QJ H 5 6 log10 Control Intensity This graph shows all the quant points IPTL only plotted control versus treated Blue is original and orange is normalised The 1 1 regulation is shown as a black dotted line VSN transforms the distribution especially for low intense data points 2 4 5 Peptide CV vs Peptide Intensity Peptide CV vs Peptide Intensity 200 n 2582 pan yi Peptide CV z 50 4 4 5 5 6 2 25 3 3 5 l Logi0 Peptide Intensity Gr
17. eceeesecessuseceseueceesaeeeeeeas 22 2 4 12 Peptide CV vs Peptide Variability cccccccssccccsseccccesecceceesececseeceeeeeeceeseuecessegeceeseaeeetes 22 2 4 13 Peptide Intensity vs Peptide Variability soseeneeeennsseensereerssreressrressrrresrreresrrresrrererreresene 23 2 4 14 Peptide Ratio vs Rank of Peptide Intensity eoesensossenserrenssrersrreessrrrsrreresrrresrreresreresene 23 2 4 15 Peptide Variability vs Peptide Angle SCOre ccccccesssccccesseccccesececeeescceeeeececeeeeceeeeeeeeeeas 24 2A I6 VME saretecstanenvn ie kowe sansa seats aes aiesnan es aes aca es os de 24 2 5 THEO AIPTL MOJU Ce ie cecs toa cai dcet tana etalk E E aaah ve santunll ENE 25 2 6 TNO O Reporter VIOGU Oizo costa dinastascaciesvacntatissnc dead deustaniucas doce cousteniccwcdssdsavabantusue dese ssepteasuancesedecs 26 2 7 Savine aN EXDOMINe Wala eicctsuts cetera tuts talventaueerentocuatnuttatatadae aac tase tak ens 26 2 8 PNRETORENCES ccs sssentvne onncoreceion pusopennseasageenecometeven shoes pucacueeah vensutndecien wus N 27 2 8 1 Cener Neaseqhoiictaseesasan sau E E A eosdvatiancanmueneenseiiansatenmeanes 28 2 8 2 PEE a T E E T A noe eee A E E 29 2 8 3 Reporter IONS resani a a a E E E 30 2 8 4 Reporter CorrecUon Mathi ranae a a A A E 31 2 8 5 a EE E E EEE TA E eas TE E E A E A E EEE E E EE EE 32 Reference n aan N panuaacamenagns sa naaaaneetannesoemunaunmnnnies 33 1 Introduction During the last decade various quant
18. er of ratios required is set during loading in the Import Data Wizard but whether to use them for protein quantification or not can be set here IPTL allows two states to be compared One is assumed Control and one Treated however this is just nomenclature and does not have to be the case Here you can set how IsobariQ should report the ratios 29 2 8 3 Reporter ions IsobariQ preferences General IPTL Reporter Ions Reporter Correction Matrix R Reporter Ion Detection MS MS tolerance Reporters Name Monoisotopic mass 115 115 108263 116 116 111618 117 117 114973 Ratios Ratios 116 115 117 115 These are the reporter specific settings however most of them are set in the Import Data Wizard greyed out here The MS MS tolerance is the match threshold for the reporters in the MS MS spectrum This is usually high resolution data and can be set low typically at 0 1 Da 30 2 8 4 Reporter Correction Matrix IsobariQ preferences General tFTL_ Reporter ions Reporter Correction Matix LR 114 1 07791 0 0689622 0 00184705 1 53194e 05 Here the computed Reporter Correction section 2 2 2 115 0 0234028 1 08706 0 0659198 0 00203391 Matrix is shown It is the inverse of the loaded text file see 116 0 0007602 0 0353113 1 08668 0 0528845 31 117 7 5814e 06 0 000352156 0 046971
19. ification techniques for proteomics have been developed including stable isotope labelling of amino acids in cell culture SILAC tandem mass tagging TMT isobaric tags for relative and absolute quantification iTRAQ and isobaric peptide termini labelling IPTL IsobariQ supports the reporter ion techniques iTRAQ TMT and others and IPTL 1 1 Reporter ion techniques iTRAQ and TMT The advantage of the reporter ion methods iTRAQ and TMT is their ability to compare up to eight different physiologic conditions within one experiment In addition unlike SILAC there is no increased complexity at the MS level while the drawbacks of iTRAQ and TMT are their relative high cost systematic dampening and that the low molecular region is not accessible to all mass spectrometers 1 2 IPTL IPTL is a recently developed isobaric quantification method for the comparison of two proteomes that is similar to the reporter ion techniques in that no increased complexity at the MS level is obtained IPTL produces several quantification points per mass spectrum which yields a robust and accurate estimate of protein abundances In addition IPTL is not hampered by the low mass cut off seen in trapping type mass spectrometers such as LTQ ion traps 2 Working with IsobariQ 2 1 Installation and licence lsobariQ can be downloaded free of charge at www biotek uio no research thiede group software and unpacked to a folder of choice using standard archiving
20. l fashion However since the distribution of protein ratios in most cases also contain regulated proteins these proteins will affect the distribution in such a way that the data are no longer normally distributed Therefore we chose to use the following approach previously described for MaxQuant for calculating the significance of a ratio in lsobariQ First all the logarithms of the normalised ratios are calculated All further calculations are performed in this log space to ensure equal treatment of up and down regulated proteins 15 Due to the fact that the distribution of ratios is not normally distributed the 15 87 percentile and the 84 13 percentile are utilised to calculate the lower and upper standard deviations of the distribution SDiower M Percentile s g7 SDupper Percentilegsi3 M Where M is the median of the log of the normalised ratios Next the z score for every ratio is calculated The z score describes how many standard deviations the observation falls from the centre of the distribution It has to be calculated with the correct SD either lower or upper depending on which side of the median the ratio is For ratios lower than the median M r zZz gt SDiower And for ratios higher than the median r M Z SDupper Where M is the median of the log of the normalised ratios and r is the log of the normalised ratio We can then ask two questions depending on the value of the ratio
21. labelled b fragments are N terminal fragments meaning that this is a fragment with a heavy label and originating from the treated sample b 6 is then the counter fragment originating from the control The correct ratio to use would then be b6 b 6 or heavy light For y6 it would have been a C terminal fragment with a light label The counter fragment y 6 would then be from the control and have the heavy label The ratio to use would still be y6 y 6 but then light heavy The ratio is calculated using the intensities of the ions If an ion can be assigned to more than one sequence fragment then it will not be used for quantification The peptide ratio is calculated as the median of all the ion ratios and the variability as the average deviation in log space Standard deviation was not chosen to describe the statistical dispersion due to its utilisation of the mean and not the median as the central tendency The average deviation is defined as variability 7 11 Where M is the median of the log transformed ratios and x is the log transformed ratio A second parameter describing the statistical dispersion is the angle score defined as the slope of a linear fit line through all ratios sorted from smallest to largest This parameter is more sensitive to variations than the average deviation and is intended to ease the finding of outlier ratios within one MS MS spectrum It is also highly influenced by the number of quantification p
22. mes 1276 lt Preferences gt lt FileIntormation infostring 1216 MSMS spectra loaded 178 proteins identified amp xa FDR 3 lt Frotein quantitated 1 numberOfPeptides 77 proteinGroup HS90A HUMAN H5902 HUMAN H5904 lt ProteinQuantification numberOtOuantitationPointstlotal 647 normalizedRatio 1 00318642596 mMsmsSpectrum annotated 0 title Elution from 44 190 to 44 190 period 0 experiment 1 MascotaAnnotation precursorNeutralMass 7886 421 boldRed 1 queryNumber 3868 hitNumbertl Hit nTlermModification _ Succinyl N term quantifiable 0 cTermModification has II Se ores lt PossibleProtein LeftFlankingResidue K ProteinName KLH10 HUMAN RightFlankingResidue ale glai lt PossibleProtein LeftFlankingResidue K ProteinName SIA10 HUMAN RightFlankingResidue ee gree tea lt RAModification mass 72 062561 position 5 name IMID 2H 4 MDHI heavy K gt lt Hit gt Please note that A LOT of data is saved and this process may take time However loading of QXML files is very quick compared to re importing the Mascot dat files lsobariQ also exports data as tab separated text files tsv files These can be imported into spread sheet applications like Excel for downstream analysis 2 8 Preferences In preferences the user can view settings selected during the loading of data in the Data Import Wizard and access settings that affect program behaviour and quantification of
23. n score Below causes IsobariQ not to load these hits Ignore Sequence Suggestions with lonscore Below causes IsobariQ to load the hits but they will not be used in quantification In the detailed Q IPTL and Q Reporter modules they will be presented as sequence suggestions to the MS MS spectrum but greyed out One exception is for IPTL when creating IPTL pairs see below section 2 2 3 where it is sufficient that only one of the two hits have ion score above this value Do Not Load Proteins with Fewer Peptides Than is a filter applied at the protein level If a protein in the dat file has fewer peptides than this value it will not be loaded Significance is the Mascot significance value when creating a peptide summary see www matrixscience com help scoring help html This parameter will highly influence the false discovery rate FDR for the identifications Maximum Proteins to Report is a protein filter Only the best scoring proteins with rank less than this value will be loaded ETD data Check this box if ETD was used as the fragmentation technique lsobariQ will then match c and z 1 fragment ions instead of b and y Import Data Wizard IPTL isobaric Peptide Termini Labeling Please select what kind of IPTL used and fill in information on the label names used in the Mascot search Standard IPTL Tryptic IPTL N terminal light label _Succinyl N term N terminal heavy label _Su
24. oints For example for a given peptide 15 ratios can be calculated varying from 1 09 to 2 98 The ratio plot below shows the ratios black dots the fitted line red and the median blue The angle O between the median and the fitted line is the angle score Ratio Ratio n 2 3 1 2 Reporter Ions This technique only uses one quantification point per MS MS spectrum and thus has no peptide variability The ratio calculation however is still not straightforward as the measured intensities of the reporters have to be corrected for isotopic overlap interference This can be achieved in IsobariQ using correction matrices as previously described by Vaudel and colleagues The matrix can be stored in a text file and imported into IsobariQ during loading See example matrices for iTRAQ and TMT in the conf folder In the iTRAQ kit the information about impurities and isotopic overlap is included For iTRAQ 4 plex it typically looks like this iTRAQ 4plex 114 0 929 59 02 iTRAQ 4plex 115 0 20 929 56 Ool iTRAQ 4plex 116 0 30 24 45 ol iTRAQ 4plex 117 o 40 924 35 ol A corresponding correction matrix reflecting this information can be created 12 92 9 2 0 0 0 5 9 92 9 3 0 0 0 2 56 92 4 4 0 0 0 1 45 92 4 C I0 The matrix has to be symmetrical The measured reporter intensities will be stored in a matrix when detected M measure and the corrected intensities are cal
25. peptide is unique if a peptide sequence has been associated with only one protein in Mascot If a peptide has been associated with more than one protein in Mascot but assigned to only one due to the Occams razor principle then the peptide is a razor peptide If Normalize rawfiles independently is checked then all the peptides identified will be normalised per raw file For gel experiments this can be very useful as a protein can be identified many places on the gel with different ratios for example due to cleavage However a certain amount of data points per raw file is needed for correct statistics so do not use this option if there are only few data points per raw file 28 2 8 2 IPTL IsobariQ preferences General IPTL Reporter ions Reporter Correction Matrix R Labels Label mass difference Quantification WM Indude peptides where Mascot could not identify both directions force found hits Control is Labeled Report Ratio N terminal light Control Treated N terminal heavy Treated Control Ok Cancel These are the IPTL specific settings however most of them are set in the Import Data Wizard greyed out here Label mass difference is the mass difference between the heavy and light IPTL labels usually 4 Da lsobariQ generates force found hits when Mascot cannot detect both forward and reverse labelling see section 2 2 3 The minimal numb
26. peptides and proteins After alterations in preferences all data should be re quantified to incorporate the changes 27 2 8 1 General IsobariQ preferences o s General IPTL Reporter Ions Reporter Correction Matrix R Loading of Data Identification Protein scoring Fragmentation MudPIT ETD Standard Spectrum Annotation and Fragment Ion Detection MS MS tolerance 0 6 Da Quantification J Require bold red from Mascot Use unique and razor peptides Use only unique peptides v Normalize rawfiles indepedently Here the user can set the scoring schema standard or MudPIT for calculating protein scores and the fragmentation techniques ETD uses c and z 1 fragments when annotating MS MS spectra Leave un checked for CID fragmentation Furthermore the threshold for matching fragment ions at the MS MS level can be set typically 0 6 Da for ion traps and lower for TOF instruments In addition there are a few settings that affect the protein quantification These settings will be applied to the individual peptides when the user clicks Quantify all proteins If Require bold red from Mascot is checked any peptide that is not bold red will not be used when calculating the protein ratio In the main view of IsobariQ the peptide quantification information will be greyed out Peptide uniqueness can be used as a filter when calculating the protein ratios The
27. re imported This is because the parameters define how IsobariQ should read the Mascot dat files and treat the results 0 Import Data Wizard _ Choose Quantification Technique 0 19 This wizard will help you load your data into Jsobarig IPTL Isobaric Peptide Termini Labeling Reporter Ions TRAO TMT Custom WOOO pees In the first frame the user chooses the kind of quantification technique that was used in the study Two techniques are supported IPTL and Reporter lons Under IPTL the methods Standard IPTL and Tryptic IPTL are supported and under Reporter lons the methods 7TRAQ TMT and Custom are supported fal Import Data Wizard Mascot Fill in information about the Mascot search Mascot Dat file C TsobariQ F024224 dat Do Not Load Sequence Suggestions With Ionscore Below 5 W Ignore Sequence Suggestions With Ionscore Below 10 Do Not Load Proteins With Fewer Peptides Than 1 Significance 0 05 Maximum Proteins To Report ETD data lt Back next gt Cancel Help In the Mascot frame the user selects the path to the Mascot dat file This file can be located either locally or on a network drive but we recommend having the file locally for performance reasons Mascot usually suggests 10 different sequences hits to every MS MS spectrum and scores them individually This is the Mascot ion score Do Not Load Sequence Suggestions with lo
28. s graph is plotted versus the rank of mean intensities it is evident that the variance is higher for low intense but actually for the 25 000 most intense points the variance is quite stable already Performing VSN on this data set yields this graph 20 Quantpoint CV vs Rank of Mean Intensity 140 n 29774 120 100 0 5000 10000 15000 20000 25000 30000 Rank of Mean Intensity The bulk of the data is not transformed only the first 5000 data points But now the variance is stable across the whole intensity range 2 4 9 Quant Point Ratio vs Mean Intensity Quantpoint Ratio vs Mean Intensity n 29774 N oy i en log10 Treated Control Oo i N i UJ logi0 Mean Intensity Graph to see the spread of quant point ratios as a function of mean intensity described above Identical to 2 4 16 2 4 10 Quant Point Ratio vs Rank of Mean Intensity Quantpoint Ratio vs Rank of Mean Intensity a log10 Treated Control am i N i UIJ 0 5000 10000 15000 20000 25000 30000 Rank of Mean Intensity 21 As previous just plotted against the rank of mean intensity A VSN version of this data yields this graph Quantpoint Ratio vs Rank of Mean Intensity UJ n 29774 N en i H log10 Treated Control oO i N i UJ TART EEn a rt yt rt rt yp rt tt yt tT tr Tt 0 5 000 10000 15000 20000 25000 30000 Rank of Mean Intensity Here we can see that for low intense data the s
29. scot query View fragments b4 of 54 Reporter Plotting and annotation png svg mage expo When IsobariQ is in Reporter mode and a protein has been double clicked the Q Reporter module will open and all the peptides belonging to that protein can be viewed and validated Detailed information like annotated MS MS spectrum possible sequence matches and quantification is available The user can interact with the quantification events something that causes IsobariQ to recalculate on the fly 2 7 Saving and Exporting Data lsobariQ saves all the data in its own format QXML This format stores all the experiment information settings peptide protein identification and quantification and all the user interactions In this way the user can go back to old data and continue validation at a later time point The QXML format is an XML based format which is also readable for humans lsobariQ does not support mzldentML at this point mainly due to its lack of support for quantification data One example of a QXML file 26 lt xml Fersion 1 0 encoding UTF 3 8 E 1DOCTYFE IsobariQML gt lt Isobarig SchemaVYersion 1 1 gt Preferences gt General useUniqueadAndRazor 1 ionscoreCutoff 10 proteinkemoveValue 0 useOnlyUnique lt IPTL nameNTerminalLight _ Succinyl N term nameLysineLight _ IMID MDHT light K naz Reporter msamsMatchiThresholdReporterions 0 1 useCorrectionMatriz 0 reporterNa
30. tal error that causes the overall distribution of ratios to deviate from 1 1 Normalisation assumes that the majority of ratios should be 1 1 One special algorithm for normalisation is the variance stabilising normalisation VSN which also reduce the variance heterogeneity found in proteomics data iii Quantify each protein individually 10 This means calculating an overall protein ratio based on all the peptides found for that protein iv Perform statistical tests Use the overall distribution of protein ratios to calculate individual protein significance using z Statistics and compute limits for up and down regulated proteins 2 3 1 Calculation of Peptide Ratios and Variability 2 3 1 1 IPTL In IPTL a peptide can have several quantification points and they are based on the identification by Mascot IPTL pair identifications consist of pairs of sequence fragments for example y5 and y 5 one belonging to the first IPTL labelling group and the other to the second labelling group When a pair has been found the ratio to use depends on the labelling of the peptide whether the fragment is a b or y or c or z 1 and finally the user s selections whether the ratio should be reported as control treated or treated control Example The control experiment is labelled with light labelled N terminals and we want to report the ratio as treated control The fragment to consider is b6 and the associated sequence is heavy N terminally
31. tatistical soread of ratios are lower than the original data above 2 4 11 Peptide CV vs Peptide Angle Score Peptide CV vs Peptide Angle Score 140 2582 Peptide CV 0 5 as 15 20 25 30 Peptide Angle Score Plot to see if the peptide CV is comparable proportional to our own metric the angle score Graph shows that the two metrics are not the same i e none is redundant 2 4 12 Peptide CV vs Peptide Variability Peptide CV vs Peptide Variability 140 n 2582 Peptide CV 0 0 2 0 4 0 6 0 8 1 Peptide Variability 22 Plot to see if the peptide CV is comparable proportional to the peptide variability average deviation The peptide CV takes into account the peptide ratio while the peptide variability does not 2 4 13 Peptide Intensity vs Peptide Variability Peptide Intensity vs Peptide Variability n n 2582 ad u ul Pyn n n log10 Peptide Intensity UJ 4 s n 0 4 0 6 0 8 1 Peptide Variability Plot to see if the peptide variability is linked to peptide intensity Note that the peptide intensity is different to the mean intensity used before see 2 4 7 The peptide intensity is the intensity calculated as the intensity sum of the fragments in the MS MS spectrum that are assigned to the sequence 2 4 14 Peptide Ratio vs Rank of Peptide Intensity Peptide Ratio vs Rank of Peptide Intensity pas u n 2582 en n log 10 Treated Control Pas
32. tion oe E Division by Channel Sum Reporter lons Only The results of the normalisation can be validated using different graphs See section 2 4 for more details 2 3 2 1 Division by Median Normalization is performed by utilising the ratios of all peptides Reporter lons or quant points IPTL The normalised peptide quant point ratio is calculated as follows Rati _ RatlOmeasured atlOnormalised M Where M is the median of all peptides quant points 2 3 2 2 Variance Stabilising Normalisation For details about VSN see previously reported data for iTRAQ or visit the homepage at www bioconductor org packages devel bioc html vsn html In brief VSN transforms the distribution of signal intensities in order to stabilise the variance across the whole intensity range The phenomenon usually seen in proteomic experiments is called heterogeneity of variance and is the dependence of the ratio on the mean signal It can be seen as a wider spread of ratios in the low intense region in a distribution plot of ratios when compared to the high intense region 2 3 2 3 Division by Channel Sum Reporter Ions Only For each reporter channel the sum of all intensities are calculated and a normalised reporter intensity for the 114 channel of iTRAQ would be calculated as follows I li14 measured 114 normalized S 114 Where J is measured intensity or rather the corrected intensity if a correction matrix is used and S is the channel s
33. um of all intensities The ratios are then calculated using the normalised intensities 14 2 3 3 Calculation of Protein Ratios and Variability 2 3 3 1 IPTL For IPTL a protein has a ratio and a normalised ratio This ratio is calculated as the median of the individual peptide ratios and the normalised ratio as the median of the individual peptide normalised ratios The protein variability is calculated as the pooled standard deviation The pooled standard deviation uses the individual peptide variability average deviation and the number of quantification points in each peptide as the basis The protein variability is then calculated as follows Xili 1 x Vi 2 i 1 Qi gt 1 protein variability Where V is the peptide variability using normalised data and Q is the number of quantification points in the peptide 2 3 3 2 Reporter Ions In reporter ion mode a protein can have several ratios and several normalised ratios each with its own variability and number of quantification points The ratios are calculated as the median of the individual peptide ratios and the variability as the average deviation see above for calculation of peptide ratio in IPTL mode 2 3 4 Calculation of Protein Ratio Significance and Limits for Up and Down Regulation If all proteins in our study have a 1 1 ratio i e no proteins were affected by our treatment all ratios will follow the central limit theorem and scatter around 1 in a log norma
Download Pdf Manuals
Related Search