Optimized Fast and Sensitive Acquisition Methods for Shotgun

Apr 27, 2012 - The recently introduced quadrupole Orbitrap (Q Exactive) tandem mass spectrometer allows fast acquisition of high-resolution higher-ene...
0 downloads 0 Views 547KB Size
Download PDF
Technical Note pubs.acs.org/jpr

Optimized Fast and Sensitive Acquisition Methods for Shotgun Proteomics on a Quadrupole Orbitrap Mass Spectrometer Christian D. Kelstrup, Clifford Young, Richard Lavallee, Michael L. Nielsen,* and Jesper V. Olsen* Department of Proteomics, Novo Nordisk Foundation Center for Protein Research, Faculty of Health Sciences, University of Copenhagen, Blegdamsvej 3b, DK-2200 Copenhagen N, Denmark S Supporting Information *

ABSTRACT: Advances in proteomics are continually driven by the introduction of new mass spectrometric instrumentation with improved performances. The recently introduced quadrupole Orbitrap (Q Exactive) tandem mass spectrometer allows fast acquisition of high-resolution higher-energy collisional dissociation (HCD) tandem mass spectra due to the parallel mode of operation, where the generation, filling, and storage of fragment ions can be performed while simultaneously measuring another ion packet in the Orbitrap mass analyzer. In this study, data-dependent acquisition methods for “fast” or “sensitive” scanning were optimized and assessed by comparing stable isotope labeled yeast proteome coverage. We discovered that speed was the most important parameter for sample loads above 125 ng, where a 95 ms HCD scanning method allowed for identification and quantification of more than 2000 yeast proteins from 1 h of analysis time. At sample loads below 125 ng, a 156 ms HCD acquisition method improved the sensitivity, mass accuracy, and quality of data and enabled us to identify 30% more proteins and peptides than the faster scanning method. A similar effect was observed when the LC gradient was extended to 2 or 3 h for the analysis of complex mammalian whole cell lysates. Using a 3 h LC gradient, the sensitive method enabled identification of more than 4000 proteins from 1 μg of tryptic HeLa digest, which was almost 200 more identifications compared to the faster scanning method. Our results demonstrate that peptide identification on a quadrupole Orbitrap is dependent on sample amounts, acquisition speed, and data quality, which emphasizes the need for acquisition methods tailored for different sample loads and analytical preferences. KEYWORDS: Q Exactive, LCMS, HCD, Orbitrap, parallel operation, data-dependent

■

only detector.8 This instrument incorporates a selection quadrupole configured with a high-efficiency C-trap and higher-energy collisional dissociation (HCD) octopole collision cell that allows both fast and sensitive generation of fragment ions.9−12 An initial report on the performance of the Q Exactive demonstrated impressive results, with more than 2500 proteins identified from a tryptic digest of HeLa cell lysate using a 90 min LC gradient.8 The coupling of a Q Exactive to an ultra high pressure liquid chromatography (UHPLC) system resulted in the identification of approximately 3900 yeast proteins from a 4 h gradient.13 Although remarkable results were achieved with the Q Exactive in these studies, investigations concerning the appropriate instrument settings for various sample conditions were not reported. In this study, we have focused on optimizing data-dependent acquisition (DDA) methods for shotgun proteomics on the Q Exactive. Since optimizing one parameter might negatively affect another, instrument optimization involves finding the most favorable balance between sensitivity (signal-to-noise),

INTRODUCTION In the postgenomic era, mass spectrometry (MS)-based proteomics has played a significant role in the large-scale analysis of proteins and greatly contributed to the understanding of gene function in biology.1 Such large-scale experiments typically involve analysis of complex protein mixtures derived from cell cultures, tissues, and/or body fluids.2,3 The complex samples are proteolytically digested, and the resulting peptide mixtures are often analyzed by nanoscale liquid chromatography (LC) coupled to a tandem mass spectrometer. In a typical setup, peptides are separated according to their hydrophobicity and then electrosprayed into the MS, where they are sequenced by tandem mass spectrometry (MS/MS).1 The ability to sequence and identify peptides from complex mixtures by MS/MS has evolved significantly over the past decade, with recent generations of mass spectrometers displaying improved sensitivity, resolution, and speed.4,5 High resolution hybrid Orbitrap mass spectrometers6,7 have become popular for proteomics analysis because of their performance capabilities, easy operation, and robustness. A new Orbitrap instrument was recently introduced (Q Exactive), which contains the Orbitrap mass analyzer as the © 2012 American Chemical Society

Received: January 10, 2012 Published: April 27, 2012 3487

dx.doi.org/10.1021/pr3000249 | J. Proteome Res. 2012, 11, 3487−3497

Journal of Proteome Research

Technical Note

Figure 1. Method design scheme for the Q Exactive. (A) Schematic of the Q Exactive describing signal-to-noise depends linearly on the ion flux and on the square root of the transient time of the Orbitrap measurement. (B) Isolation of the 262.6, 524.3 m/z MRFA or 1322 m/z peak of the Ultramark calibration solution was used to measure the relative increase in injection time which was inverted to give the dependence of the isolation window on the ion flux. (C) A fixed cycle time is measured as a function of the two methods where the number of fragmentation events is varied (top N methods). An example of results obtained for resolution of 17 500 (red) and 35 000 (green) is used, and the result of linear regression of the measurements is shown (D) Each MS/MS event in the Q Exactive consists of an “isolation/fragmentation etc.” step where injection time (blue) is a major part and an Orbitrap measurement step where the transient time (orange) is the major contribution. These occur in parallel (symbolized with a red arrow) indicating the transfer of ions into the Orbitrap but with different additional time required between scans (black). The scatter plot depicts timing measurements of these additional timings (black) and the relationship between the transient time (orange) and the injection time (blue). The line labeled “line of parallel acquisition” indicates an optimal parallel isolation/fragmentation and Orbitrap measurement as a function of resolution (transient time) and injection time. Two optimal settings are identified and marked by stars named S (sensitive) and F (fast) in the figure. The solid horizontal lines indicate possible resolution settings with the red dotted line showing a hypothetical setting for a 32 ms transient. The thin vertical dotted lines indicate the settings possible for the injection time.

increased injection time is therefore required to maintain sensitivity. Improving the data quality by using longer injection times or increased resolution comes at the cost of higher measurement times, which decreases the overall acquisition speed. This can be mitigated by a feature of the Q Exactive to accumulate ions in parallel to the acquisition of the transient.8 However, this also complicates further optimization as the optimal resolution now depends on the ideal ion injection time. Thus, a systematic evaluation of relevant instrumental parameters is important toward achieving greater numbers of peptide and protein identifications on the Q Exactive. We designed two acquisition methods, one optimized for sensitivity (“sensitive” method) and the other optimized for speed (“fast” method). The performance of these methods was measured by analyzing a dilution series of stable isotope labeling by amino acids in cell culture (SILAC)-labeled yeast lysates.18 The sensitive 156 ms HCD method of the Q Exactive performed better for restricted sample amounts (30% improved sequencing speed should result in more sequenced peptides and consequently more peptides identified. We compared the number of identified peptides reported from each experimental setup and discovered the LTQ Orbitrap Velos managed to identify on average 4491 unique peptides within the 60 min gradient, while the fast and sensitive Q Exactive methods identified on average 10 784 and 8572 unique peptides, respectively (Table 3). Overlapping the uniquely identified peptides from each method revealed that 87% of unique peptides identified by the LTQ Orbitrap Velos were also identified in the two Q Exactive methods (Figure 3D). Overall, the sensitive Q Exactive method identified 48% more unique peptides (an increase from 4491 to 8572) than the LTQ Orbitrap Velos, while the fast Q Exactive method identified an additional 26% (from 8572 to 10 784). It was concluded that both Q Exactive methods target the same peptide species as the LTQ Orbitrap Velos and that the increased sequencing speed produced both more sequenced and identified peptides. We investigated the overall MS/MS data quality acquired from the three experimental setups. The overall mass accuracy of the HCD spectra acquired from the LTQ Orbitrap Velos and the fast Q Exactive method were similar, with a standard deviation on all acquired fragment ions of 4.0 and 4.3 ppm, respectively (Figure 4A). This is perhaps not surprising since the recalculated HCD resolution for the LTQ Orbitrap Velos was 10 500 when compared to a HCD resolution of 17 500 in the fast Q Exactive method (both resolutions at m/z 200). In comparison, the standard deviation on HCD spectra acquired with the sensitive Q Exactive method was only 2.8 ppm, which is a substantial improvement in fragment ion precision and is most likely due to the 2-fold higher resolution HCD spectra (resolution 35 000 at m/z 200). Although the MaxQuant database search settings were not altered between the methods, a stricter filter could be implemented on the fragment mass tolerance for the sensitive method due to the increased fragment ion accuracy. Besides the observed improvements in mass accuracy, acquiring data for longer injection times and at higher resolution should produce higher fragment ion abundances and increased sensitivity. We calculated the total signal abundance for all acquired fragmentation scans by multiplying the total ion current (TIC) with the injection time from the individual methods (Figure 4B). An almost 3-fold increase in signal abundance was reported for the sensitive Q Exactive method when compared to the fast Q Exactive method. For the LTQ Orbitrap Velos, a 6-fold decrease in signal abundance was observed compared to the fast Q Exactive. It should be noted that the internal software scaling of signal intensities is different between the two generations of Orbitrap mass spectrometers, which most likely is the main reason for the observed differences. The altered signal abundances may also reflect instrument differences with regard to how the fragmentation spectra were acquired, as the LTQ Orbitrap Velos performed fragmentations with a fixed target value, while the Q Exactive operated with fixed ion injection times. As a consequence, the latter often produces a larger dynamic range of fragment ion abundances in HCD spectra, so fragment ion abundances between instrument platforms are not directly comparable. We also investigated the database search results generated by the individual instrumental setups by searching the acquired RAW files with the Andromeda search engine.22 The

Figure 4. Qualitative comparison of acquisition methods. Different qualitative parameters are investigated for the fast (red) and sensitive (green) Q Exactive method and the reference (blue) LTQ Orbitrap Velos method. For all box plots, the whiskers indicate 1.5 times the interquartile range. (A) A box plot displays the median centered fragment mass accuracy for all annotated peaks in the MS/MS spectra. (B) A box plot displays the fragment ion count distributions between the different methods. The fragment ion count for each identified MS/ MS spectrum is calculated as the total ion current for the spectrum multiplied by the injection time. (C) Box plots are shown for the Andromeda score distributions found for each of the methods. A 3494

dx.doi.org/10.1021/pr3000249 | J. Proteome Res. 2012, 11, 3487−3497

Journal of Proteome Research

Technical Note

Figure 4. continued higher score indicates a higher amount of annotated peaks. (D) Quantization accuracy is displayed for each method by plotting the measured SILAC ratio for each identification event as a function of its intensity. Density distributions of the SILAC ratio are shown on top of each plot. (E) Histograms for proteins identified in these studies have been binned on the basis of the estimated copy number of each protein measured previously in a reference data set.32 Proteins where no reference information was found are depicted in a separate column.

probability-based Andromeda peptide score distribution for all identified peptides showed that HCD spectra generated by the Q Exactive in the sensitive setting produced a median peptide score of 95, slightly above the median peptide score of 91 obtained by the LTQ Orbitrap Velos. In contrast, the fast Q Exactive method provided the lowest peptide scores with an overall median score of 76 (Figure 4C). These results reflect once again the difference in data quality between the utilized methods, with the sensitive method generating higher quality MS/MS spectra, while the fast method is optimized for speed. An important aspect of proteomic experiments is the ability to perform relative quantitation.33 Since all samples were acquired using a SILAC-labeled yeast mixture, it was possible to compare the quantitative performance of all three setups. We found that the overall SILAC quantitation accuracy was only minimally affected by the two Q Exactive methods when compared to the LTQ Orbitrap Velos (Figure 4D), despite the ability of the Q Exactive to sample and identify peptides of lower abundance (Figure 4E). In absolute numbers, the standard deviations of the log 2 converted protein ratios were 0.78, 0.74, and 0.68 for the Q Exactive fast, sensitive, and LTQ Orbitrap Velos methods, respectively. The minor differences reflect a complex interplay between increases in quantitation accuracy through faster cycle times (increased number of sampling points across the eluting peptide peaks) versus decreases resulting from the improved coverage of lower abundance peptides. This relationship is demonstrated in Figure 4D, where ratio measurements display a larger spread at lower intensity values for both Q Exactive methods. Technological improvements in chromatography and MS sensitivity will be required to address these issues. Interestingly, it seems that both Q-Exactive sensitive and fast methods cover a protein abundance dynamic range of 5 orders of magnitude in single runs (Figure 4D), while the LTQ Orbitrap Velos is limited to 4 orders of magnitude in dynamic range. Finally, we performed an analysis of a tryptic digest of HeLa whole cell lysates in order to evaluate the performance of the two established Q Exactive methods on a complex human sample using longer LC gradients. As depicted in Figure 5A, the number of MS/MS scans increase with gradient duration, with the fast method producing the highest number of scans. The ability to identify MS/MS scans was found to decrease with increased gradient duration, with the sensitive method having the highest identification rates (Figure 5B). This demonstrates that shorter gradients should be used in conjunction with fast methods, whereas with longer gradients the sensitive method yields the best results (Figure 5C). From an optimization standpoint, the trade-off between speed and sensitivity changes with the gradient length, with a longer gradient being equivalent to a dilution of the analyzed sample. Noteworthy, the peptide identifications from a 3 h gradient analyzed with the sensitive method collapsed into more than 4000 proteins,

Figure 5. Gradient length dependence of fast and sensitive acquisition methods. One microgram of HeLa tryptic peptides were analyzed by 1, 2, and 3 h gradients. (A) The number of MS/MS events increased with the gradient length and is higher for the fast (red) method than the sensitive (green) method. (B) Identification rate at a fixed 1% FDR for the sensitive (green) method was higher than the fast (red) method. (C) The number of peptides counted on the sequence level was highest for the sensitive (green) method for the longest gradient, while the fast (red) method resulted in more hits for shorter gradients. (D) The number of protein groups were highest for the sensitive (green) method for the longest gradient, while the fast (red) method resulted in more hits for shorter gradients. (E) Venn diagram of proteins identified in duplicates from the 2 h gradient with the fast (red) method and the sensitive (green) method. (F) Venn diagram of proteins identified in the 1 and 3 h gradients with the fast (red) method and the sensitive (green) method.

3495

dx.doi.org/10.1021/pr3000249 | J. Proteome Res. 2012, 11, 3487−3497

Journal of Proteome Research

Technical Note

proteomic analysis, as just a few years ago it required almost two days of MS measurements to cover a similar depth of the yeast proteome.19 We believe that the future challenges of proteomics technologies are in the area of improving the sensitivity of the Orbitrap. If the raw sensitivity of the Orbitrap analyzers remains unchanged, then the ion flux into the instrument will become the limiting factor. An alternative approach to increase sensitivity involves the use of UHPLC systems to improve chromatographic separations.

which is almost 200 proteins more than the fast method (Figure 5D). Replica analysis shows a large degree of overlap between protein identification, despite the random nature of DDA (Figure 5E). As expected, the shorter gradients produced protein identifications that to a very large degree were contained within the identifications from the longer gradients (Figure 5F).

■

CONCLUSIONS In this technical note, we have established two different DDA methods for the Q Exactive. The methods have been optimized for the parallel operation mode, which is required to maximize the ability of the quadrupole Orbitrap mass spectrometer to sequence and identify peptides from complex mixtures. We have designed the methods aimed at either fast or sensitive operation. At amounts above 125 ng of yeast peptides, both protocols were pushed to their respective scanning speed limit, which ultimately penalized the slower scanning speed of the sensitive method compared to the fast scanning method. Conversely, at lower amounts, fewer fragment scans were acquired by the sensitive method, and the penalty for spending more time on each fragment scan was largely offset by the improvement in data quality. Likewise, we find a similar effect when extending the LC gradient, with the sensitive method outperforming the fast scanning method when analyzing 1 μg of a complex HeLa lysate for 3 h, which resulted in more than 4000 proteins identified. Our analysis reveals general guidelines for which method to use and how numerous instrument parameters can be assessed. Briefly, if the system is not using all the available fragment scans specified in a top N acquisition method, more time can be spent on each MS/MS scan by using a more sensitive method. Alternatively, if the spectra are of insufficient quality to enable confident peptide identifications, more time can be spent on each spectrum to increase the signal-to-noise. The parallel operation of the Q Exactive is powerful, but the instrumental settings have to be chosen correctly in order for the parallel ion injection to function optimally. Having established the two most favorable instrument methods, we compared their analytical performance using a dilution series of SILAC labeled yeast lysates. Although the major analytical differences between the fast and sensitive Q Exactive methods were revealed, it also highlighted the compromise between MS/MS acquisition speed and spectral quality typically encountered in proteomic experiments. Clearly for studies where spectrum quality is important, such as the analysis of low abundance samples or when accurate PTM site assignment is desired,34 the sensitive method would be the preferred protocol to use. However, for complex and abundant samples where extensive profiling is essential, the fast operation of the Q Exactive would be more appropriate. Our analyses demonstrate that the faster scanning method is superior at sample amounts above 125 ng, whereas the sensitive method performs better at lower sample loads. To highlight the recent advances in mass spectrometers, we mapped all identified yeast proteins identified by the presented experimental setups to their cellular copy numbers (Figure 4E).32 While the LTQ Orbitrap Velos analysis produced relatively poor coverage of the low copy proteins, considerably more of these proteins were uniquely identified by the Q Exactive methods. Interestingly, the Q Exactive was able to identify and quantify more than half of the yeast proteome in only 1 h of analysis. This is another significant step forward in

■

ASSOCIATED CONTENT

S Supporting Information *

Figure S1 contains method details of the fast and sensitive Q Exactive methods. Figures S2 and S3 contain annotated spectra according to JPR guidelines from yeast and HeLa. Figure S4 describes details of the technical timing experiments. Figure S5 depicts XIC for dilution series of yeast analyzed by the fast and sensitive Q Exactive methods. Figure S6 depicts XIC for the replica 250 ng of yeast samples analyzed on the LTQ Orbitrap Velos and by the fast and sensitive Q Exactive methods. Figure S7 depicts details of the replica 250 ng samples analyzed on the LTQ Orbitrap Velos and by the fast and sensitive Q Exactive methods. Table S1 contains a list of identified proteins. Table S2 contains details from the HeLa dilution series analyzed by three different acquisition methods. This material is available free of charge via the Internet at http://pubs.acs.org/.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]; [email protected]. dk. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank all members of the Department for Proteomics at the Center for Protein Research for fruitful discussions and inputs to the manuscript. We are especially grateful to Jon W. Poulsen, Dorte Bekker-Jensen, and Chiara Francavilla for providing the lysates. The Center for Protein Research is partly supported by a generous donation from the Novo Nordisk Foundation. This work was supported by the seventh framework program of the European Union (Contract No. 262067- PRIME-XS). Figure 1A is reprinted with the permission of ThermoFisher Scientific.

■

REFERENCES

(1) Aebersold, R.; Mann, M. Mass spectrometry-based proteomics. Nature 2003, 422, 198−207. (2) Cox, J.; Mann, M. Quantitative, high-resolution proteomics for data-driven systems biology. Annu. Rev. Biochem. 2011, 80, 273−299. (3) Cravatt, B. F.; Simon, G. M.; Yates, J. R., 3rd The biological impact of mass-spectrometrybased proteomics. Nature 2007, 450, 991−1000. (4) Domon, B.; Aebersold, R. Mass spectrometry and protein analysis. Science 2006, 312, 212−217. (5) Mann, M.; Kelleher, N. L. Precision proteomics: the case for high resolution and high mass accuracy. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 18132−18138. (6) Makarov, A.; Denisov, E.; Kholomeev, A.; Balschun, W.; Lange, O.; Strupat, K.; Horning, S. Performance evaluation of a hybrid linear ion trap/orbitrap mass spectrometer. Anal. Chem. 2006, 78, 2113− 2120.

3496

dx.doi.org/10.1021/pr3000249 | J. Proteome Res. 2012, 11, 3487−3497

Journal of Proteome Research

Technical Note

(7) Olsen, J. V.; Schwartz, J. C.; Griep-Raming, J.; Nielsen, M. L.; Damoc, E.; Denisov, E.; Lange, O.; Remes, P.; Taylor, D.; Splendore, M.; Wouters, E. R.; Senko, M.; Makarov, A.; Mann, M.; Horning, S. A dual pressure linear ion trap Orbitrap instrument with very high sequencing speed. Mol. Cell. Proteomics 2009, 8, 2759−2769. (8) Michalski, A.; Damoc, E.; Hauschild, J.-P.; Lange, O.; Wieghaus, A.; Makarov, A.; Nagaraj, N.; Cox, J.; Mann, M.; Horning, S. Mass spectrometry-based proteomics using Q Exactive, a high-performance benchtop quadrupole Orbitrap mass spectrometer. Mol. Cell. Proteomics 2011, 10, M111.011015−. (9) Koecher, T.; Pichler, P.; Schutzbier, M.; Stingl, C.; Kaul, A.; Teucher, N.; Hasenfuss, G.; Penninger, J. M.; Mechtler, K. High precision quantitative proteomics using iTRAQ on an LTQ Orbitrap: a new mass spectrometric method combining the benefits of all. J. Proteome Res. 2009, 8, 4743−4752. (10) Olsen, J. V.; Macek, B.; Lange, O.; Makarov, A.; Horning, S.; Mann, M. Higher-energy C-trap dissociation for peptide modification analysis. Nat. Methods 2007, 4, 709−712. (11) Frese, C. K.; Altelaar, A. F. M.; Hennrich, M. L.; Nolting, D.; Zeller, M.; Griep-Raming, J.; Heck, A. J. R.; Mohammed, S. Improved peptide identification by targeted fragmentation using CID, HCD and ETD on an LTQ-Orbitrap Velos. J. Proteome Res. 2011, 10, 2377− 2388. (12) Zhang, Y.; Ficarro, S. B.; Li, S.; Marto, J. A. Optimized Orbitrap HCD for quantitative analysis of phosphopeptides. J. Am. Soc. Mass Spectrom. 2009, 20, 1425−1434. (13) Nagaraj, N.; Kulak, N. A.; Cox, J.; Neuhaus, N.; Mayr, K.; Hoerning, O.; Vorm, O.; Mann, M. Systems-wide perturbation analysis with near complete coverage of the yeast proteome by single-shot UHPLC runs on a bench-top Orbitrap. Mol. Cell. Proteomics 2012, 11, M111.013722. (14) Makarov, A.; Denisov, E.; Lange, O.; Horning, S. Dynamic range of mass accuracy in LTQ Orbitrap hybrid mass spectrometer. J. Am. Soc. Mass Spectrom. 2006, 17, 977−982. (15) Marshall, A. G. Theoretical signal-to-noise ratio and mass resolution in Fourier transform ion cyclotron resonance mass spectrometry. Anal. Chem. 1979, 51, 1710−1714. (16) Michalski, A.; Cox, J.; Mann, M. More than 100,000 detectable peptide species elute in single shotgun proteomics runs but the majority is inaccessible to data-dependent LC−MS/MS. J. Proteome Res. 2011, 10, 1785−1793. (17) Savitski, M. M.; Sweetman, G.; Askenazi, M.; Marto, J. A.; Lang, M.; Zinn, N.; Bantscheff, M. Delayed fragmentation and optimized isolation width settings for improvement of protein identification and accuracy of isobaric mass tag quantification on Orbitraptype mass spectrometers. Anal. Chem. 2011, 83, 8959−8967. (18) Ong, S.-E.; Blagoev, B.; Kratchmarova, I.; 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, 376−386. (19) de Godoy, L. M. F.; Olsen, J. V.; de Souza, G. A.; Li, G.; Mortensen, P.; Mann, M. Status of complete proteome analysis by mass spectrometry: SILAC labeled yeast as a model system. Genome Biol. 2006, 7, R50. (20) Nielsen, M. L.; Vermeulen, M.; Bonaldi, T.; Cox, J.; Moroder, L.; Mann, M. Iodoacetamideinduced artifact mimics ubiquitination in mass spectrometry. Nat. Methods 2008, 5, 459−460. (21) Rappsilber, J.; Ishihama, Y.; Mann, M. Stop and go extraction tips for matrix-assisted laser desorption/ionization, nanoelectrospray, and LC/MS sample pretreatment in proteomics. Anal. Chem. 2003, 75, 663−670. (22) Cox, J.; Neuhauser, N.; Michalski, A.; Scheltema, R. A.; Olsen, J. V.; Mann, M. Andromeda: a peptide search engine integrated into the MaxQuant environment. J. Proteome Res. 2011, 10, 1794−1805. (23) 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, 1367−1372.

(24) Cox, J.; Mann, M. Computational principles of determining and improving mass precision and accuracy for proteome measurements in an Orbitrap. J. Am. Soc. Mass Spectrom. 2009, 20, 1477−1485. (25) Olsen, J. V.; de Godoy, L. M. F.; Li, G.; Macek, B.; Mortensen, P.; Pesch, R.; Makarov, A.; Lange, O.; Horning, S.; Mann, M. Parts per million mass accuracy on an Orbitrap mass spectrometer via lock mass injection into a C-trap. Mol. Cell. Proteomics 2005, 4, 2010−2021. (26) R Development Core Team R: A Language and Environment for Statistical Computing; R Foundation for Statistical Computing: Vienna, Austria, 2010. (27) Michalski, A.; Damoc, E.; Lange, O.; Denisov, E.; Nolting, D.; Mueller, M.; Viner, R.; Schwartz, J.; Remes, P.; Belford, M.; Dunyach, J.-J.; Cox, J.; Horning, S.; Mann, M.; Makarov, A. Ultra high resolution linear ion trap Orbitrap mass spectrometer (Orbitrap Elite) facilitates top down LC MS/MS and versatile peptide fragmentation modes. Mol. Cell. Proteomics 2012, 11, O111.013698. (28) Makarov, A.; Denisov, E.; Lange, O. Performance evaluation of a high-field Orbitrap mass analyzer. J. Am. Soc. Mass Spectrom. 2009, 20, 1391−1396. (29) Elias, J. E.; Gygi, S. P. Target-decoy search strategy for increased confidence in large-scale protein identifications by mass spectrometry. Nat. Methods 2007, 4, 207−214. (30) Miyamoto, K.; Hara, T.; Kobayashi, H.; Morisaka, H.; Tokuda, D.; Horie, K.; Koduki, K.; Makino, S.; Núñez, O.; Yang, C.; Kawabe, T.; Ikegami, T.; Takubo, H.; Ishihama, Y.; Tanaka, N. High-effciency liquid chromatographic separation utilizing long monolithic silica capillary columns. Anal. Chem. 2008, 80, 8741−8750. (31) Thakur, S. S.; Geiger, T.; Chatterjee, B.; Bandilla, P.; Fröhlich, F.; Cox, J.; Mann, M. Deep and highly sensitive proteome coverage by LC−MS/MS without prefractionation. Mol. Cell. Proteomics 2011, 10, M110.003699. (32) Ghaemmaghami, S.; Huh, W.-K.; Bower, K.; Howson, R. W.; Belle, A.; Dephoure, N.; O’Shea, E. K.; Weissman, J. S. Global analysis of protein expression in yeast. Nature 2003, 425, 737−741. (33) Bantscheff, M.; Schirle, M.; Sweetman, G.; Rick, J.; Kuster, B. Quantitative mass spectrometry in proteomics: a critical review. Anal. Bioanal. Chem. 2007, 389, 1017−1031. (34) Kelstrup, C. D.; Hekmat, O.; Francavilla, C.; Olsen, J. V. Pinpointing phosphorylation sites: Quantitative filtering and a novel site-specific x-ion fragment. J. Proteome Res. 2011, 10, 2937−2948.

3497

dx.doi.org/10.1021/pr3000249 | J. Proteome Res. 2012, 11, 3487−3497