The One Hour Yeast Proteome*

We describe the comprehensive analysis of the yeast proteome in just over one hour of optimized analysis. We achieve this expedited proteome characterization with improved sample preparation, chromatographic separations, and by using a new Orbitrap hybrid mass spectrometer equipped with a mass filter, a collision cell, a high-field Orbitrap analyzer, and, finally, a dual cell linear ion trap analyzer (Q-OT-qIT, Orbitrap Fusion). This system offers high MS2 acquisition speed of 20 Hz and detects up to 19 peptide sequences within a single second of operation. Over a 1.3 h chromatographic method, the Q-OT-qIT hybrid collected an average of 13,447 MS1 and 80,460 MS2 scans (per run) to produce 43,400 (x̄) peptide spectral matches and 34,255 (x̄) peptides with unique amino acid sequences (1% false discovery rate (FDR)). On average, each one hour analysis achieved detection of 3,977 proteins (1% FDR). We conclude that further improvements in mass spectrometer scan rate could render comprehensive analysis of the human proteome within a few hours.

The ability to measure differences in protein expression has become key to understanding biological phenomena (1,2). Owing to cost, speed, and accessibility, transcriptomic analysis is often used as a proteomic proxy (3,4). That said, mRNA is a genetic intermediary and cannot inform on the myriad of post-translational regulation processes (5)(6)(7). For the past decade considerable effort has been invested in maturing proteomic technology to deliver information at a rate and cost commensurate to transcriptomic technologies.
Historically yeast, with its 6600 open reading frames, has been the preferred proteomic technology test-bed (8). In 2003, Weissmann and colleagues measured approximate ex-pression levels of each yeast gene using either GFP or TAP tags (9). This seminal work established that ϳ4500 proteins are expressed during log-phase yeast growth. Subsequent mass spectrometry-based studies have confirmed this early estimate (9 -12). With this knowledge, we hereby define comprehensive proteome analysis as an experiment that detects ϳ90% of the expressed proteome (Ն 4000 proteins for yeast). Note others have used the term "nearly complete" for this purpose; we posit that comprehensive has identical meaning (i.e. including many, most, or all things) (13).
Initial MS-based proteomic analyses of yeast, each identifying up to a few hundred proteins, were conducted using a variety of separation and MS technologies (14 - 16). Yates and co-workers reported the first large-scale yeast proteome study in 2001 with the identification of 1483 proteins following ϳ 68 h of mass spectral analysis, i.e. 0.4 proteins were identified per minute (17). Their method-two dimensional chromatography coupled with tandem mass spectrometryhas provided a template for large-scale protein analysis for the past decade (18 -20). By incorporating an offline first dimension of separation with more extensive fractionation (80 versus 15) Gygi et al. expanded on this work in 2003 (21). That said, the modest increase in identified proteins (1504) required 135 h of analysis, reducing the protein per minute count to 0.2. Armed with a faster hybrid mass spectrometer capable of accurate mass measurement, Mann and colleagues achieved detection of 2003 yeast proteins in an impressive 48 h (0.7 proteins/minute) in 2006 (22).
From these three pioneering studies we begin to see the impact of mass spectrometer acquisition rate on the depth and rate of proteome analysis. The most recent application of such technology to the yeast proteome, however, The Mann work used a hybrid linear ion trap-ion cyclotron resonance Fourier transform instrument (LTQ-FT) that delivered MS 2 scans at a rate of ϳ650 ms (23). The earlier studies, i.e. Yates and Gygi, relied on the considerably slower scanning (1-3 s/scan) three-dimensional ion trap technology. In 2008, using the novel Orbitrap hybrid mass spectrometer, Mann and colleagues reported on the first comprehensive analysis of the yeast proteome by identifying nearly 4000 proteins (10). Extensive fractionation (24) and triplicate analysis of each fraction rendered the study a considerable time investment at ϳ144 analysis hours (0.5 proteins/minute). In 2010 our group achieved similar comprehensive analysis, but improved sequence coverage, using fractionation and multiple proteases (24). That work, however, required even longer analysis time (0.2 proteins/min).
And that was the state-of-the-art as recently as three years ago. Doubtless we, the proteomics community, had achieved one momentous goal-comprehensive coverage of the yeast proteome. Still, obtaining this depth was not routine as it mandated days of MS analysis and a considerable amount of expert labor. In 2012, with new, even faster scanning, quadrupole-Orbitrap technology (Q-OT, Q-Exactive), Mann and colleagues dispatched the concept of fractionation, improved the quality of sample preparation, and placed emphasis on higher quality online separations (25). With their streamlined method they achieved detection of just over 3900 yeast proteins following four hours of MS analysis. Even more impressive this strategy translated to the identification of 16.3 proteins per minute-a 33-fold improvement over the next best comprehensive study. This success was a remarkable achievement and illustrates that comprehensive proteomic technology can indeed be executed in a time efficient manner.
Time-of-flight hybrid systems, of course, can deliver very high MS 2 acquisition rates, up to 100 Hz in some reports. In 2011, Muddiman and colleagues reported yeast proteome analysis using a quadrupole-TOF system (i.e. TripleTOF) operating at a much lower rate (20 Hz) MS 2 scan rate (26). Even at this reduced rate, only 16% of the spectra were mapped to unique sequences and 1112 unique proteins identified. Because of reduced MS 2 spectral quality (i.e. low signal-tonoise, S/N), even fewer unique peptide identifications were achieved at higher MS 2 acquisition rates. Other studies using TOF technologies report similar results (27,28). For maximal proteome depth, we conclude that increased scan speed must not come at the cost of reduced spectral quality. Recently, a new Orbitrap hybrid mass spectrometer having a mass filter, a collision cell, a high-field Orbitrap analyzer, and, finally, a dual cell linear ion trap analyzer was described (Q-OT-qIT, Orbitrap Fusion) (29,30). This system offers high MS 2 acquisition speed of 20 Hz-double that of the Q-OT system used by Mann and colleagues. We postulated that this fresh system, with its fast scan rate, could provide comprehensive proteome analysis in record time. To maximize performance we developed an optimized cellular lysis approach, employed trypsin digestion, and used dimethyl sulfoxide (DMSO, 5%) as an LC additive to increase abundance of acidic peptides and unify charge state (31,32). Using this novel system we report the comprehensive analysis of the yeast proteome (4002 with 1% FDR) following 1.3 h of nLC-MS 2 analysis (70 min gradient). These experiments delivered an extraordinary 67 proteins per minute and demonstrate that complete analysis of the yeast proteome can be routinely performed in approximately one hour.

EXPERIMENTAL PROCEDURES
Yeast Culture and Lysis-Saccharomyces cerevisiae strain BY4741 was grown in yeast extract peptone dextrose media (1% yeast ex-tract, 2% peptone, 2% dextrose). Four liters of media was divided between four two-liter flasks and inoculated with a starter culture (OD 600 ϭ 2.58). Cells were allowed to propagate for ϳ12 generations (20 h) to an OD 600 ϳ2 (average of 2.18). The cells were harvested by centrifugation at 5000 rpm for 5 min, supernatant decanted, resuspended in chilled NanoPure water and all pellets were pooled together. The cells were washed two more times and centrifuged for the final pelleting at 5000 rpm for 10 min. A pellet corresponding to 5% of the total cells grown, was resuspended in lysis buffer composed of 50 mM Tris pH8, 8 M urea, 75 mM sodium chloride, 100 mM sodium butyrate, protease (Roche) and phosphatase inhibitor tablet (Roche). Yeast cells were lysed by glass bead milling (Retsch). Briefly, 2 ml of acid washed glass beads were combined with 2.5 ml of resuspended yeast cells in a stainless steel container and shaken 8 times at 30 hz for 4 min with a 1 min rest in between.
Protein Digestion-Lysate protein concentration was measured by BCA (Thermo Pierce). Protein was reduced by addition of 5 mM dithiothreitol and incubated for 45 min at 55°C. The mixture was cooled to room temperature, followed by alkylation of free thiols by addition of 15 mM iodoacetamide in the dark for 30 min. The alkylation reaction was quenched with 5 mM dithiothreitol. Urea concentration was diluted to 1.5 M with 50 mM Tris pH 8.0. Proteolytic digestion was performed by addition of Trypsin (Promega, Madison, WI), 1:50 enzyme to protein ratio, and incubated at ambient temperature overnight. An additional 1:50 bolus of trypsin was added in the morning and incubated at ambient temperature for 1 h. The digestion was quenched by addition of TFA and desalted over a tC18 Sep-Pak (Waters, Milford, MA).
nLC-MS 2 Analysis-Reversed phase columns were prepared inhouse. Briefly, a 75-360 m inner-outer diameter bare-fused silica capillary, with a laser pulled electrospray tip, was packed with 1.7 m diameter, 130 Å pore size, Bridged Ethylene Hybrid C18 particles (Waters) to a final length of 35 cm. The column was installed on a nanoAcquity UPLC (Waters) using a stainless steel ultra-high pressure union formatted for 360 m outer diameter columns (IDEX) and heated to 60°C for all runs. Mobile phase buffer A was composed of water, 0.2% formic acid, and 5% DMSO. Mobile phase B was composed of acetonitrile, 0.2% formic acid, and 5% DMSO. Samples were loaded onto the column for 12 min at 0.35 l/min. Mobile phase B increases to 4% in the first 0.1 min then to 12% B at 32 min, 22% B at 60 min, and 30% B at 70 min, followed by a 5 min wash at 70% B and a 20 min re-equilibration at 0%B.
Eluting peptide cations were converted to gas-phase ions by electrospray ionization and analyzed on a Thermo Orbitrap Fusion (Q-OT-qIT, Thermo). Survey scans of peptide precursors from 300 to 1500 m/z were performed at 60K resolution (at 200 m/z) with a 5 ϫ 10 5 ion count target. Tandem MS was performed by isolation at 0.7 Th with the quadrupole, HCD fragmentation with normalized collision energy of 30, and rapid scan MS analysis in the ion trap. The MS 2 ion count target was set to 10 4 and the max injection time was 35 ms. Only those precursors with charge state 2-6 were sampled for MS 2 . The dynamic exclusion duration was set to 45 s with a 10 ppm tolerance around the selected precursor and its isotopes. Monoisotopic precursor selection was turned on. The instrument was run in top speed mode with 5 s cycles, meaning the instrument would continuously perform MS 2 events until the list of nonexcluded precursors diminishes to zero or 5 s, whichever is shorter. Elite runs were performed with Survey scans of peptide precursors from 300 to 1500 m/z 60K resolution (at 200 m/z) with a 1 ϫ 10 6 ion count target. Tandem MS was performed by isolation at 1.8 Th with the ion-trap, CAD fragmentation with normalized collision energy of 35, and rapid scan MS analysis in the ion trap. The data dependent top 20 precursors were selected for MS 2 . MS 2 ion count target was set to 5 ϫ 10 3 and the max injection time was 125 ms. Only those precursors with charge state ϩ2 or higher were sampled for MS 2 . The dynamic exclusion duration was set to 40 s with a 10 ppm tolerance around the selected precursor and its isotopes. Monoisotopic precursor selection was turned on.
Data Analysis-The raw data was processed using Proteome Discoverer (version 1.4.0.288, Thermo Fischer Scientific). MS 2 spectra were searched with SEQUEST engine against a database of 6632 yeast open reading frames (ORFs) 1 (www.yeastgenome.com, February 3, 2011) (33). Peptides were generated from a tryptic digestion with up to two missed cleavages, carbamidomethylation of cysteines as fixed modifications, and oxidation of methionines and protein N-terminal acetylation as variable modifications. Precursor mass tolerance was 20 ppm and product ions were searched at 0.35 Da tolerances. Peptide spectral matches (PSM) were validated using percolator based on q-values at a 1% FDR (34). With proteome Discoverer, peptide identifications were grouped into proteins according to the law of parsimony and filtered to 1% FDR (35). For cumulative protein group identification, PSMs passing the FDR were exported to a text file and processed by a modified version of Protein Hoarder (version 2.4.1) (36). The PSMs were iteratively processed in successive 1 min windows and grouped into proteins using the law of parsimony at a 1% FDR.

RESULTS
Considerable gains in the depth and rate of proteomic analysis have been realized over the past decade (vide supra). These improved results stem from routine use of high mass accuracy and resolution, but also from a steady increase in MS 2 acquisition rate. In the decade spanning the seminal Yates publication in 2001 and the single-shot proteome work of Mann et al. in 2012, MS 2 sampling rates rose from ϳ0.75 Hz to nearly 10 Hz (17,25). Here we report on an even newer generation of mass spectrometer that comprises a mass re-solving quadrupole, Orbitrap, collision cell, and linear ion trap (Q-OT-qIT, Fusion, Fig. 1) (29,30). In this system, MS acquisition rate is not only boosted by the presence of a very fast scanning dual cell linear ion trap, but also by a control environment having multiple, independent processing units. The new system is equipped with a sophisticated control system that parallelizes the processes of ion injection, precursor isolation, fragmentation, and mass analysis to achieve a ϳ2ϫ boost in acquisition rates. We reasoned that this Q-OT-qIT configuration, with its 20 Hz MS 2 acquisition rate, could afford a considerable gain for rapid, whole proteome analysis.
The Q-OT-qIT System-To test this hypothesis we began by performing a parametric evaluation using a complex mixture of yeast tryptic peptides eluted into the system over a 70 min gradient. We examined several settings including collisional activation mode (i.e. HCD or trap CAD), MS 1 resolution, collision energy, maximum inject time, and dynamic exclusion settings. Detailed plots highlighting these results are included in Supplemental Information. Briefly, we found that MS 2 analysis using HCD followed by ion trap mass analysis (low-res HCD; 80,626 MS 2 events with 33,127 unique PSMs) generated more identifications compared with ion trap CAD with ion trap mass analysis (CAD; 75,973 MS 2 events with 31,820 unique PSMs). This is not surprising as HCD tends to offer more random backbone fragmentation and, with the Q-OT-qIT geometry, can be accomplished slightly faster. Operation of the system with an MS 1 resolving power setting of 60,000 (@ m/z 200) afforded a 20% increase in detected unique peptides over 15,000 resolving power (supplemental Fig. S1). We conclude the boosted resolving power elevates precursor signal-to-noise (S/N) ratios, allowing for improved selection of low abundance precursors, and can potentially separate otherwise unresolved precursors so that multiple MS 2 events can be acquired. Of course, in this scenario such closely spaced precursors would be co-isolated (0.7 m/z isolation width); however, the selected precursor m/z annotated in the MS 2 scan would be different and would facilitate identification from a chimeric MS 2 scan. Increasing MS 1 resolving power above 60,000 did not provide any apparent benefit for increased identifications. Thirty-five milliseconds was the optimal maximum injection time (supplemental Fig. S2). Decreasing the maximum injection time to 30 msec, or increasing it to 45 msec caused a 10% decrease in peptide identifications. We found only slight variations in peptide identifications among dynamic exclusion settings of 30, 45, and 60 s (supplemental Fig. S3). Quadrupole isolation widths from 0.5 to 1.5 m/z were examined, the best results were achieved at a value of 0.7 m/z (supplemental Fig. S4). MS 1 and MS 2 automatic gain control (AGC) target values of 500,000 and 7000, respectively, produced the maximum number of peptide identifications (supplemental Figs. S5 and S6).
Lysis, Chromatography, and Additives-Yeast cell lysis is a critical step in achieving comprehensive proteome detection and must be executed with care. Detergents, such as SDS, or bead beating are typical approaches for yeast lysis (37)(38)(39). The SDS method, used in the Mann yeast studies, has produced excellent results, but requires removal of the detergent before MS sampling. The bead beating method mixes glass beads and yeast cells in a buffer slurry, which is shaken for three, 1-4 min cycles, at 30 hz. This approach, however, can be too gentle to sufficiently lyse the yeast and, in our hands, does not efficiently extract all proteins. We aimed to avoid use of detergents and investigated a more vigorous bead beating procedure. By simply extending the number of cycles to eight (4 min each) we achieved considerably improved results (supplemental Fig. S7). Finally, we note increased identifications when lysates were not cleared of insoluble material. Zubarev and colleagues recently reported similar findings for mammalian cell culture samples (40).
Previous single-shot yeast proteome analyses utilized long capillary LC columns (50 cm) and lengthy gradients (240 min) (25,40,41) We aimed to achieve comparable or better coverage using a much shorter gradient (70 min). We found that capillary LC columns of 30 cm packed with 1.7 m BEH particles (Waters Corporation) operating at flow rates of 350 -375 nL/minute offered consistent elution across the one hour gradient. To accommodate this flow rate, a home-built column heater was maintained at a temperature of 60°C throughout the separation. Sample was loaded directly onto the column to avoid losses.
Recent work by Kuster et al. described that addition of ϳ5% DMSO to the mobile phase solvents boosted precursor S/N, providing up to a 20% increase in protein identifications (31,32). We tested addition of DMSO to our chromatography solvents by comparing the number of yeast peptide and protein identifications obtained over our 70 min gradient either with or without DMSO. In our hands the presence of DMSO increased the average precursor signal, from ϳ2.8 ϫ 10 7 to 4.8 ϫ 10 7 (arbitrary units) and increased the total ion current by 170% (supplemental Fig. S8). This amplified signal afforded a 9% increase in unique peptide identifications and 5% more proteins. We conclude that DMSO can indeed improve performance, with no apparent downside, and included it for all subsequent experiments.
Whole Yeast Proteome Analysis-To test our supposition that the faster scanning Q-OT-qIT hybrid could deliver comprehensive yeast proteome analysis within ϳ1 h, we sequentially analyzed trypsin-digested, yeast cell lysate in quintuplicate. Each replicate began by loading ϳ1.4 g of sample and followed by analysis over a 70 min gradient. Allowing for sample loading, column washing, and equilibration, the five consecutive analyses consumed ϳ8 h; however, actual instrument acquisition times were ϳ1.3 h per experiment. As anticipated, the Q-OT-qIT hybrid posted a considerable number of scans: on average 13,447 MS 1 and 80,460 MS 2 events per run. As a point of reference, state-of-the-art analysis in 2003, by Gygi et al., recorded 162,000 MS 2 events following 135 h of MS operation. The Q-OT-qIT hybrid delivered this number of scans in two and a half hours! Next, we analyzed the yeast sample on the most recently introduced quadrupole linear ion trap Orbitrap system (i.e. qIT-OT or Orbitrap Elite) using the identical chromatographic conditions (42). That mass spectrometer produced only about a quarter of the MS 1 scans, as compared with the Q-OT-qIT (3635), and half the MS 2 events (39,447). Fig. 2 presents a series of MS 2 scans acquired by the Q-OT-qIT MS over a 1 s period. In this example, 22 precursors were selected for MS 2 analysis from MS 1 scan #59,211. All 22 product ion spectra were acquired, individually, within 1 s and are presented in the lower portion of Fig. 2. Nineteen of these 22 scans were mapped to sequence (1% FDR). On database searching, each one hour experiment (Q-OT-qIT system) produced 43,400 (x ) peptide spectral matches (PSMs) and 34,255 (x ) peptides with unique amino acid sequences (1% FDR, Table I). Batched analysis of the five experiments yielded 47,624 unique peptides. In each analysis, over half of the 80,460 MS 2 scans were mapped to sequence (54%). Despite the swift Q-OT-qIT scan rate (ϳ20 Hz), we conclude the system routinely delivers spectra of high quality.
On average, each of the quintuplicate analyses achieved detection of 3,977 proteins (1% FDR), 13.5% (538, x ) of which originate from single peptides (Fig. 3A). Combination of the data reduces single peptide proteins to 460 while deepening coverage to 4395 protein groups. 3643 of these proteins (83%) were present in all five 1-h experiments and 3853 (88%) were found in four of five (Fig. 3B). Median sequence coverage was 18.4% and 23.7% for the individual and combined experiments, respectively (Fig. 3C), with a median of seven identified peptide sequences per protein. Yeast contains ϳ800 dubious ORFs. These ORFs, which are believed not to encode a corresponding protein, are commonly used to verify the FDR of proteomic data sets. For the one hour experiments, between three and eight dubious ORFs were identified per dataset (Table I), confirming these data are indeed well below the 1% FDR threshold.
To directly contrast the performance of the Q-OT-qIT hybrid to the most recent comprehensive yeast analysis we analyzed the same samples using a 240 min gradient. This longer method mimics the 2012 study of Mann and colleagues (vide infra). With the extended gradient conditions the Q-OT-qIT system identi-  To estimate the dynamic range of the one hour experiments we compared our mass spectrometry-based identifications to those mapped by either tandem affinity (TAP) or green fluorescent protein (GFP) tagging experiments. From our one hour experiment data we identify 89% of proteins for which there is abundance data, including 73% of proteins present at less than 125 copies/cell (Fig. 3D) (9). We also note detection of 886 proteins lacking abundance data. Next, we benchmarked the dynamic range of our global analysis to a recent multiple reaction monitoring (MRM) study. There, Aebersold et al. targeted 152 yeast proteins, spanning the full concentration range including several proteins never observed in public proteomic data sets, using synthetic peptides and triple quadrupole MRM technology (43). A single one hour Q-OT-qIT experiment, batched analysis of all five 1-h experiments, and our four hour analysis netted 122, 133, and 132 of the 152 Aebersold targets, respectively. The Aebersold work detected 137 of these protein targets likely following dozens of MRM experiments. We conclude that our one hour method, using the Q-OT-qIT hybrid, provides dynamic range and sensitivity comparable to state-of-the-art MRM studies, but with whole proteome depth. We also note that the system, with its quadrupole mass filter, offers considerable promise for parallel reaction monitoring (44,45). DISCUSSION The data presented above provides a deep view of the yeast proteome. Equally important is that this depth is achieved within an unprecedented time-scale. To understand how these gains were realized we plotted the number of identified peptide sequences per second as a function of elution time (Fig. 4A) for both the Q-OT-qIT (red) and the previous generation qIT-OT (black) hybrids. Remarkably, this plot reveals that the Q-OT-qIT hybrid (Orbitrap Fusion) routinely identifies ϳ8 peptides per second with occasions where up to 19 peptide sequences are detected in a single 1 s window. These stunning metrics are approximately double that achieved by the qIT-OT (Orbitrap Elite). Further, the speed of the Q-OT-qIT allows for deeper MS 2 sampling of the MS 1 . The mean precursor depth sampled by the Q-OT-qIT is the 349 th most abundant m/z peak in the MS 1 scan whereas the qIT-OT system achieves an average depth of only 202 (Fig. 4B). The Q-OT-qIT frequently samples m/z peaks that are the 800th or weaker in intensity rank.
To further examine the effect of Q-OT-qIT scan speed we plotted the cumulative number of unique peptide sequence identifications as a function of retention time (Fig. 4C, I/L ambiguity removed). Almost linearly across the 70 min gradient and wash period the Q-OT-qIT (solid red) accumulates unique peptide identifications at a rate considerably faster than the qIT-OT (solid black)-8.3 versus 4.3 unique sequences per second, respectively (linear fit between 10 and 80 min). Use of a four hour gradient on this same system slightly reduces the slope (3.7 unique sequences per second, linear fit between 10 and 200 min, dotted red), but allows for considerably deeper analysis and again outpaces the qIT-OT system (2.3 unique sequences per second, dotted black). For reference we plotted these same data for the 2012 Mann et al. study using the Q-OT hybrid (Q-Exactive, dotted blue). That system posts the shallowest slope (1.7 unique sequences per second) and approximately half the number of unique peptide sequences as compared with the Q-OT-qIT (Orbitrap Fusion) for the same analysis time. Note that Mann used a Lys-C digest in this work. Panel D of Fig. 4 compares unique protein identification rates for the same data sets. Again the Q-OT-qIT system is the top performer, even when comparing the one hour analysis (solid red) to four hour experiments on either of the other systems (dotted black and blue).
Here we described new mass spectrometer technology that is capable of achieving comprehensive yeast proteome coverage within an unprecedented time-scale. Doubtless over the past decade many improvements to sample preparation, chromatography, and MS hardware have contributed to making this achievement possible. Among all these, we attribute increased mass spectrometer scan speed as the primary reason for the acceleration in proteome analysis speed and depth. Fig. 5 illustrates the pace of protein identifications for several large-scale yeast proteomic analyses as a function of the mass spectrometer MS 2 scan rate. Note the rapid ascent in protein identification rates scales correlates with increasing MS 2 scan rate.
The correlation depicted in Fig. 5 was somewhat surprising to us as we expected that ionization suppression of lower abundance peptides would become increasingly dominant as complex peptide mixtures comprising whole proteomes are separated over shorter gradients -i.e. from four to one hour (46,47). In other words, as the separation duration of the online chromatography is compressed, increased co-elution must occur. With increased co-elution one might expect that, regardless of the MS speed or sensitivity, ionization suppression would prevent a considerable fraction of peptides from becoming gas-phase ions -a requisite for MS detection. The results shown here refute this hypothesis and confirm that further improvements in MS sensitivity and speed will continue to reduce whole proteome analysis time, most likely to less than one hour for relatively simple proteomes like yeast.
Finally, we conclude that comprehensive analysis of mammalian proteomes within several hours is now within our technical reach. Consider that recent estimates suggest between 10,000 and 12,000 proteins are expressed at any given time for human cells in culture (48 -50). That is only approximately three to four times the complexity of the yeast proteome. Thus, our current efforts are aimed at achieving comprehensive coverage of mammalian system within just a few hours of analysis. Looking forward, one more doubling of MS 2 acquisition rate, i.e. from 20 to 40 Hz, has potential to deliver detection of the whole human proteome in just one to two FIG. 4. Analytical metrics of yeast proteome analysis using the Q-OT-qIT (Fusion) as compared with qIT-OT (Orbitrap Elite) and Q-OT (Q-Exactive) hybrids. The Q-OT-qIT (panel A, red) achieves identification of up to 19 peptides per second as compared with 10 with the qIT-OT system (A, black). Peak depth is likewise considerably higher on account of the faster MS 2 scanning rate of the Q-OT-qIT system (B). C, plots the pace of unique yeast peptide identifications for the three instruments. For the one hour analysis, the Q-OT-qIT posts almost twice as many unique peptide identifications as compared with the qIT-OT. Similar data, except for unique proteins, is shown in D. hours. And, given the history and rate of MS innovation, such capability is likely only a few years away.