Enhanced Sample Multiplexing-Based Targeted Proteomics with Intelligent Data Acquisition

Abstract

Targeted proteomics has been playing an increasingly important role in hypothesis-driven protein research and clinical biomarker discovery. We previously created a workflow, Tomahto, to enable real-time targeted pathway proteomics assays using two-dimensional multiplexing technology. Coupled with the TMT 11-plex reagent, hundreds of proteins of interest from up to 11 samples can be targeted and accurately quantified in a single-shot experiment with remarkable sensitivity. However, room remains to further improve the sensitivity, accuracy, and throughput, especially for targeted studies demanding a high peptide-level success rate. Here, bearing in mind the goal to improve peptide-level targeting, we introduce several new functionalities in Tomahto, featuring the integration of gas-phase fractionation using the FAIMS device, an accompanying software program (TomahtoPrimer) to customize fragmentation for each peptide target, and support for higher multiplexing capacity with the latest TMTpro reagent. We demonstrate that adding these features to the Tomahto platform significantly improves overall success rate from 89% to 98% in a single 60 min targeted assay of 290 peptides across human cell lines, while boosting quantitative accuracy via reducing TMT reporter ion interference.

Graphical Abstract

INTRODUCTION

Mass spectrometry-based quantitative proteomics has become one of the major approaches to interrogate protein alteration upon perturbation.^1,2 While shotgun proteomics enables global discovery of new leads without requiring any prior hypothesis,^3,4 targeted proteomics is known for its sensitivity and robustness when a group of proteins of interest (e.g., biomarkers, signaling pathways) must be assayed reproducibly across a great number of biological samples.^5–7 Classic targeted proteomics through multiple reaction monitoring (MRM) or selected reaction monitoring (SRM) require tedious transition curation and are subject to many variables, limiting transfer across platforms.^7,8 While many targets can be multiplexed in one experiment, the fact that only a single sample can be analyzed at once limits overall throughput, and the challenge becomes more prominent when hundreds or thousands of samples must be assessed in a timely fashion.

In addition to target multiplexing, sample multiplexing provides an orthogonal option to enhance throughput in targeted proteomics. Sample multiplexing techniques via the use of isobaric labeling reagents have evolved iteratively and now allow up to 21 samples to be analyzed simultaneously in a single assay.^9–12 As one of the commercially available reagents, tandem mass tag (TMT) has been widely adopted in quantitative proteomics and is proven advantageous not just for throughput, but also for reducing variability incurred during sample preparation and analysis.^13–16 A strategy, termed triggered by offset, multiplexed, accurate-mass, high-resolution, and absolute quantification (TOMAHAQ), incorporating isotopically labeled synthetic trigger peptides (STPs) with TMT11plex-based sample multiplexing is developed by Erickson et al, which obviates need for peptide detection prior to quantification and thereby substantially extends the limit of quantification for low abundance targets.^17,18 STPs can be either synthesized using heavy isotope-labeled amino acids, or by labeling regular peptides using the super heavy TMT (shTMT) reagent.^17–19 Noting that the TOMAHAQ method is complex and can become cumbersome when the target list is more than dozens, we subsequently built a user-friendly workflow, termed Tomahto, to enable real-time instrument control and decision making via an application programming interface (API)-based algorithm.²⁰ Tomahto significantly reduces the amount of prior knowledge needed to conduct an experiment targeting hundreds of proteins²⁰ and has expanded the scope of its applicability to nontryptic peptides such as Aβ peptides²¹ and even unconventional spliced isoforms when integrated with proteogenomics.²²

While Tomahto is readily applicable to a wide spectrum of hypothesis-driven protein research, it relies on multiple peptides for a protein target to be fail safe, and the peptide-level success rate might be only moderate. However, growing interest across a spectrum of subjects, including assessing proteoforms,²² quantifying HLA peptides in antigen presentation,²³ and measuring cell signaling mediated by various post-translational modifications (PTMs),²⁴ can only be achieved by peptide-level quantification. In light of this challenge, we note that a few technical advancements have been made lately, which we hypothesize could further improve the platform. The TMTpro reagent increases sample multiplexing capacity from the original TMT11-plex to 18-plex.^10,11,25 Corresponding to the shTMT, a super heavy TMTpro reagent (shTMTpro) is later generated with a mass shift of +9 Da relative to TMTpro.²⁶ High-field asymmetric-waveform ion mobility mass spectrometry (FAIMS) offers an orthogonal dimension in signal enrichment and improves accuracy in isobaric tag-based quantification.^27–29 In addition, most existing TMT-based workflows, targeted or shotgun, assume a universal fragmentation strategy, exemplified by the classic collision-induced dissociation (CID).^20,29–32 However, peptides possess distinct physiochemical properties, and their behavior varies dramatically when subjected to same fragmentation strategy.³³ This challenge has prompted the development of many orthogonal fragmentation approaches,³⁴ including HCD,³⁵ ETD,³⁶ and UVPD³⁷ among others and many of their variants.³³ We reason that being able to determine and use an optimal fragmentation scheme for each individual target peptide is critical in targeted proteomics, especially when very few surrogate peptides are of great biological interest. This challenge becomes more prominent when assessing peptides bearing labile post-translational modifications (e.g., phosphorylation,^38,39 glycosylation^40,41).

Here we introduce three new functionalities into the Tomahto platform (Figure 1). First, TMTpro reagents are now supported by Tomahto so that up to 18 samples can be measured simultaneously. However, we caution against using the TMTpro135 channel due to potential interference from the shTMTpro reporter, which has an identical mass. Second, FAIMS-based gas-phase fractionation is integrated and its benefit to sensitivity and accuracy is highlighted. Last, we provide the community a software tool, TomahtoPrimer, to determine optimal fragmentation settings for each target peptide and demonstrate its ability to further improve the targeted experiment. We showcase the newly added functionalities with an experiment targeting 290 peptides with a single-shot 60 min analysis. Overall, we are able to increase the peptide-level quantification success rate from 89% to 98% while enhancing quantitative accuracy.

Figure 1.

Sample multiplexing-based targeted proteomics enabled by intelligent data acquisition. New functionalities are implemented in the instrument control software Tomahto, which dictates data collection via the instrument application program interface (iAPI). These new features include (1) integration of gas-phase fractionation using FAIMS device, (2) a software tool, TomahtoPrimer, to support customizing fragmentation for each peptide target, and (3) support for TMTpro18-plex reagents for higher multiplexing. In a Tomahto experiment, synthetic trigger peptides (STPs) are labeled with superheavy TMTpro reagent.²⁶ The STPs are first analyzed in a priming run to detemine optimal fragmentation scheme for each STP. STPs are then spiked in a multiplexed sample labeled using the TMTpro18-plex reagent. Tomahto monitors the elution of STPs and collects trigger MS2 scans using primed fragmentation setting for each target. Once the identity of the STP is confirmed, Tomahto triggers a target MS2 scan by m/z offset using the specific compensation voltage (CV) value in which the STP is detected to confirm the presence of the corresponding endogenous target, therefore obviating the detection of its precursor and extending the dynamic range. Subsequently, a SPS-MS3 scan is collected through Tomahto to acquire quantitative information across up to 18 multiplexed samples.

METHODS

Reagents.

Reagents for tissue culture, including DMEM, fetal bovine serum (FBS), penicillin/streptomycin, and phosphate-buffered saline (PBS) were obtained from Gibco. Mass spectrometry-grade trypsin and Lys-C protease were purchased from ThermoFisher Scientific and Wako, respectively. Isobaric TMT reagents and the BCA protein concentration assay kit were purchased from ThermoFisher Scientific. Empore-C18 material for in-house StageTips was acquired from CDS and Sep-Pak cartridges were purchased from Waters. Synthetic trigger peptides (Table S1) were obtained from JPT Peptide Technologies.

Algorithm and Software Package.

Tomahto was written in C# in the.NET Framework (v4.6.2). It has three modules, namely, data acquisition, real-time data visualization, and data analysis. Tomahto utilizes the Fusion instrument API (freely available from ThermoFisher Scientific, https://github.com/thermofisherlsms/iapi). Tomahto is freely available via a free user license for the Orbitrap Tribrid mass spectrometer platforms (i.e., Lumos, Eclipse, and Ascend). TomahtoPrimer was written in C# using.NET 6.0 and is distributed together with Tomahto. Access can be requested through https://gygi.hms.harvard.edu/software.html.

Preparation of Cell Line Sample.

Human RPE1 (#CRL-4000), U2OS (#HTB-96), HCT116 (#CCL-247), and HEK293T (#CRL-3216) cells were purchased from the American Type Culture Collection and grown in DMEM supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin until 80% confluent. Cells were washed twice with ice cold PBS, pelleted, and stored at −80 °C until use. Cell pellets were processed as described previously.⁴² In brief, cells were lysed by resuspension in lysis buffer, followed by 10 passes through a 21-gauge syringe. Lysates were reduced with 5 mM tris(2-carboxyethyl)phosphine (15 min, room temperature [rt]) and alkylated with 10 mM iodoacetamide (30 min, rt in the dark). Excess iodoacetamide was quenched with 10 mM dithiothreitol (15 min, rt). Proteins were isolated by chloroform methanol precipitation, subsequently resuspended in 200 mM EPPS pH 8.5 (~1 mg/mL) and digested first with LysC for 12 h at rt. shaking on a vortexer followed by a 6-h digestion at 37 °C with trypsin. Protein digests were aliquoted to the desired concentrations and labeled directly with separate TMT channels. The labeled peptides were then mixed and desalted on a SepPak prior to basic pH fractionation or LC-MS analysis.

Preparation of shTMTpro-Labeled Synthetic Trigger Peptides (STPs).

The mixture of STPs (50 pmol of each) was resuspended in 200 mM EPPS (pH = 8.5). HPLC grade acetonitrile was added to a final concentration of 30% (v/v) followed by adding shTMTpro reagents (Thermo Fisher Scientific; catalog no. A52040) at a ratio of 2:1 (shTMTpro:-Peptide; w/w). The reaction proceeded at room temperature for 1 h before quenching with a final volume of 0.5% hydroxylamine (Sigma). The peptides were then vacuum-dried and purified using a 50 mg SepPak. The peptides were then reconstituted in 1% formic acid in water and were now ready for MS analysis.

Liquid Chromatography–Mass Spectrometry Analysis.

All experiments were performed on an Orbitrap Eclipse mass spectrometer coupled to a Proxeon NanoLC1200 UHPLC. FAIMS was used for all experiments, unless otherwise specified. The 100 μm capillary column was packed with 30 cm of Accucore 150 resin (2.6 μm, 150 Å;ThermoFisher Scientific). Mobile phases were 5% acetonitrile and 0.125% formic acid (A) and 95% acetonitrile and 0.125% formic acid (B). Tomahto experiment without FAIMS was performed as reported previously.²⁰ Briefly, STPs were spiked in TMTpro-labeled multiplexed samples. Twenty-five fmol of STPs and 250 ng were loaded and analyzed using a 60 min method. Tomahto employs the real-time filtering mechanism as previously described.³⁰ Tomahto listened to each collected MS1 scan. When a precursor ion matches a potential trigger peptide (±5 ppm mass accuracy; matched charge state; minimal intensity of 5e⁴), the software will insert up to four scans to be collected by the Orbitrap. (1) Tomahto prompts insertion of an Orbitrap MS2 scan with the trigger peptide’s precursor m/z. Once collected, a real-time peak matching strategy (RTPM) is used to confirm the identity of the trigger peptide (must match >6 fragment peaks within ±10 ppm). (2) If successful, Tomahto prompts the insertion of an Orbitrap MS2 scan using the target peptide as the precursor m/z value. The target peak m/z is a mixture of multiplexed endogenous peptides. At the same time, the MS2 fragment ions and their intensities from the trigger peptide are stored in memory as a template library spectrum. After collection, the target MS2 scan is used to confirm that the target peptide is present at levels sufficient for detection. This is accomplished via RTPM where fragment ions must be present in the spectrum (±10 ppm) and rank ordered by intensity from the trigger MS2. SPS fragment ions are now selected from this scan. Only b- and y-type ions are considered for selection provided they have a TMT modification. SPS candidates are required to match the fragmentation pattern of the stored library spectrum, meaning fragment ratios relative to the highest fragment were within ±50% of that in the stored spectrum. In addition, each SPS candidate undergoes a purity filter of 0.5 (at least 50% of the signal attributed to the fragment ion within a 3 m/z window) to be included in the final list. (3) Upon confirmation of target peptide presence and successful selection of SPS ions, Tomahto next triggers an ion trap SPS-MS3 prescan (max injection time of 10 ms). This is used to quickly estimate the signal strength for the TMT reporter ions. This estimate is used to set the lengthy injection times needed for the SPS-MS3 scan detected in the Orbitrap. 4) Following the prescan, Tomahto prompts the insertion of the analytical SPS-MS3 quantification scan.

For Tomahto experiments using the FAIMS device, MS1 scans were collected alternating between 3 different compensation voltages (−40, −60, and −80 V). Tomahto monitors the elution of STPs and triggers subsequent MS2 and SPS-MS3 scans by using the CV value with which the STP is detected.

For Tomahto experiments with optimized fragmentation schemes, a priming run was performed using only shTMTpro-labeled STPs. Ten MS2 scans were collected on each STP using different fragmentation settings, namely 5 CID and 5 HCD scans with collision energies of 25, 27, 30, 32, and 35, respectively. The priming run was loaded in TomahtoPrimer, and the fragmentation setting that generated the most SPS-eligible signal was selected. A target peptide list with an optimized fragmentation setting was exported and used for the subsequent Tomahto analysis.

Data Analysis.

For quantification, TMT reporter ion signal-to-noise (SN) values were extracted from MS3 scans and those with a summed SN < 160 across sixteen channels were removed from final data set. SPS-MS3 scans with SPS ion isolation specificity < 0.5 were also removed. Column normalization was performed to correct for different protein loading in each channel. Lastly, for each peptide signal-to-noise measurements were summed and then normalized to 100 across the multiplexed samples yielding a “relative abundance” measurement. Results were further analyzed using R (ver. 4.3.2)⁴³ in RStudio.⁴⁴

RESULTS AND DISCUSSION

Targeted proteomics is becoming increasingly important in proteomics research⁴⁵ and clinical studies⁴⁶ to evaluate specific sets of proteins by measuring multiple surrogate peptides. Targeting multiple surrogate peptides per protein ensures robustness and increases the likelihood of successful protein quantification, even though the success rate in quantifying each individual peptide is often moderate.^17,20 However, mounting evidence suggests that the ability to interrogate specific peptides is crucial to a wide spectrum of biomedical research. For example, spliced isoforms may only differ by a single amino acid yet hold great biological implication in health and disease.^22,47,48 Moreover, post-translational modification (PTM) plays a vital role in mediating cellular homeostasis, whose assessment relies on quantifying the peptide bearing the specific PTM.²⁴,⁴⁹,⁵⁰ The pressing demand for a dependable platform has spurred us to enhance Tomahto even further (Figure 1) driven by our overarching ambition to advance research across the spectrum of disciplines that necessitate reliable peptide-level targeting.

We first incorporated TMTpro labeling into the workflow, with the aim of further increasing the sample multiplexing capacity. We created an assay with 290 peptide targets and spiked the synthetic trigger peptides (STPs) into a 16-plex sample consisting of biological quadruplicates of four human cell lines (Figure 2A). 250 ng of TMTpro-labeled endogenous peptides and 25 fmol of each shMTpro-labeled STP were analyzed using a 60 min method with Tomahto. The STPs were labeled with the superheavy TMTpro reagent (shTMTpro) which is +9 Da heavier than the TMTpro 18-plex reagents. The TMTpro135 reporter possesses an identical number of heavy isotopes and thereby an identical mass as the shTMTpro reporter. Although the shTMTpro-labeled STP is outside the quadrupole isolation window (0.5 m/z), given that usually the STP mixture is spiked in at a much higher level relative to the endogenous target, a small fraction of the shTMTpro-labeled background matrix or isotopic impurity in the shTMTpro reagent can present an appreciable interfering signal in the TMTpro135 reporter derived from endogenous peptide targets. The 16-plex sample did not include TMTpro135. However, we observed that the TMTpro135 signal varied from being absent to dominant due to interference (Figure 2B–D). We calculated the relative abundance of each channel across all of the SPS-MS3 scans collected in the Tomahto experiment. TMTpro135 spanned a larger abundance range compared to the 16 channels used for labeling, whereas TMTpro134c—the other unused channel—was essentially absent as expected (Figure 2B). Similar observation was also made when an 18-plex sample was analyzed (Supplemental Figure S1). Therefore, the TMTpro135 reagent should be used with caution in Tomahto assays.

Figure 2.

Tomahto assay using TMTpro reagents. (A) a panel of 290 synthetic trigger peptides was labeled with super heavy TMTpro (shTMTpro) reagent and spiked into a 16-plex sample consisting of biological quadruplicates of four human cell lines. (B) Relative reporter abundance in all quantified SPS-MS3 scans (N = 2717 from triplicate Tomahto analyses). (C) Example SPS-MS3 scan of peptide IVVVTAGVR. No interference in the TMTpro135 channel is present. (D). Example SPS-MS3 scan of peptide TLSFGSDLNYATR. Interference in the TMTpro135 channel is present.

We then integrated FAIMS-based gas-phase fractionation into Tomahto, based on previous observations that FAIMS can effectively removes singly charged ions^27,51,52 and therefore enrich signals from multiply charged peptides. During Tomahto analysis, the instrument alternates between different CV values. Tomahto monitors elution of STPs and, upon potential detection within a narrow mass tolerance window (±5 ppm), instructs the instrument to collect a trigger MS2 scan for confirmation using the same FAIMS CV with which it is detected.

We benchmarked Tomahto in triplicate with and without the FAIMS interface. After a quantification filter was applied, experiments employing FAIMS resulted in an average of 274.3 quantified peptides, achieving a 95% success rate. In contrast, without FAIMS, Tomahto quantified 259 peptides (Figure 3A). We attribute the significant increase to the ability of FAIMS to remove singly charged ions which conversely enriches multiply charged targets (Supplemental Figure S2).^28,52 We then calculated ratios of protein abundance from the other three cell lines relative to RPE1. While most of the quantifications showed strong agreement between the two data sets (Pearson R = 0.95), we observed a slight inclination toward values measured by FAIMS at higher fold changes (Figure 3B; Table S2). This trend is consistent with a previous observation indicating that FAIMS effectively mitigates ratio compression induced by coisolated ions.^27,54 This reduction in ratio compression can be best exemplified by the peptide GYFFLDER derived from the PDZ and LIM domain proteins 4 (PDLIM4). An 8.7× difference was measured between U2OS and HCT116 without FAIMS, whereas with FAIMS, a 167× difference was observed (Figure 3C). This difference closely parallels the 134× difference in gene expression, as reported by the Cancer Cell Line Encyclopedia (Figure 3D).⁵³

Figure 3.

Integration of FAIMS-based gas-phase fractionation with Tomahto improves sensitivity and accuracy. Tomahto assay targeting the 290 peptides was performed with and without FAIMS in triplicate using a 60 min method. (A) Tomahto assay with FAIMS significantly boosted the number of successfully quantified peptides (p value = 0.0027). Bars represent mean ± SD from triplicate experiments. (B) Protein abundance measurements between data collected with and without FAIMS are compared for U2OS, HEK293T, and HCT116 relative to RPE1. The dotted line is y = x and the orange line represents linear regression. (C) Example peptide measurement. Instead of an 8.7× difference without FAIMS, Tomahto measured a 167× change with FAIMS for the peptide GYFFLDER. (D) PDLIM4 gene expression data by CCLE suggested a 134× difference between U2OS and HCT116.⁵³ TPM: transcripts per million.

Next, we hypothesized that the ability to customize fragmentation schemes for each target peptide could further enhance Tomahto. While CID is commonly employed for TMT-based quantitative proteomics,^3,20 the distinct characteristics of each peptide refute a universal fragmentation method and thus necessitate peptide-specific optimization to achieve optimal peptide quantification.³³ A priming run was first conducted with each shTMTpro-labeled STP analyzed with 10 different fragmentation settings (Figure 4A). Here we tested 5 normalized collision energies (NCE) using CID and HCD, respectively, on the 290 target STPs. We created a software tool, TomahtoPrimer, to process the resulting data. TomahtoPrimer matches each scan in the priming run using the same filtering mechanism that Tomahto applies in real-time. The quantitative SPS-MS3 scan in a Tomahto experiment relies on the availability of fragment ions from the preceding MS2 scan that can be further selected and fragmented to produce TMT reporter ions (i.e., SPS-eligible ions). SPS-eligible ions include only fragment ions that bear TMT labeling and fall within the user-defined m/z range. For tryptic peptides ending in lysine, both the b- and y-ion series are considered, whereas the y-ion series is not considered for peptides ending in arginine. The b1 and y1 ions are excluded due to low specificity. Fragments below 400 m/z, where interfering peaks are prevalent, are also excluded. Precursor mass exclusion is set to 50 m/z below and 3 m/z above the precursor m/z to avoid selecting precursors and neutral losses. Therefore, TomahtoPrimer prioritizes fragmentation settings that produce the most intense SPS-eligible signal. We primed all 290 target peptides and ~45% of the targets preferred HCD over CID as the fragmentation mode (Table S1). For example, the ten scans collected on peptide DTNGSQFFITTVK were ranked by TomahtoPrimer (Figure 4B). Contrary to CID at a default NCE of 35, which produced only 7 SPS-eligible fragments, HCD at an NCE of 25 more than doubled this count to 16 (Figure 4B, C). More remarkably, the percentage of eligible SPS intensity tripled from 14.32% to 45.01% by switching the fragmentation setting (Figure 4C).

Figure 4.

TomahtoPrimer allows the determination of optimal fragmentation parameters for each target peptide. (A) A priming run was performed to test different fragmentation settings on the shTMTpro-labeled STP. Data were analyzed by TomahtoPrimer to select the best setting. A target list with optimized fragmentation settings was exported and used in the subsequent Tomahto experiment. (B) Ten fragmentation settings were ranked for the peptide DTNGSQFFITTVK. The best setting based on eligible SPS ions was highlighted in green, while the default setting was indicated in gray. (C) Representative MS2 scans for peptide DTNGSQFFITTVK using the default (CID, NCE = 35%) and the primed fragmentation (HCD, NCE = 25%).

We then adopted the primed target list and performed Tomahto assays. These assays averaged 10 more quantified peptides compared to the default CID, achieving a 98% success rate within a single 60 min analysis (Figure 5A). To highlight the sensitivity of the Tomahto assay, we compared these data to a deep-fractionated data set collected on the latest Orbitrap Ascend mass spectrometer.⁵⁵ The same 16-plex sample was fractionated into 12 basic pH fractions and analyzed using a 90 min method and the real-time search-based MS3 acquisition (RTS-MS3).⁵⁵ Tomahto averaged 284 quantified peptides, whereas the deep fractionated data set, despite spending 18 h and quantifying more than 86,900 peptides from 9,100 proteins, only captured 222 of the 290 target peptides (Figure 5B). One of the peptides that were quantified only using the primed HCD fragmentation was peptide FYDSWESTVK from the lysine deficient protein kinase 1 (WNK1). Although the overall number of SPS-eligible fragment ions were similar between primed and default settings (12 vs 10), the total and the percentage of intensity attributable to these fragments doubled from 17% to 36% (Figure 5C). During the following Tomahto experiments, despite the trigger MS2 looking identical, a dominant interfering peak at 623 m/z was present in the target MS2 scan (Figure 5D). This peak was likely derived from coisolated ions, and as a result all fragment peaks from the true endogenous target were suppressed. Though present at a low level, the target was successfully quantified due to the primed fragmentation. Upon inspecting the Tomahto data without using the primed fragmentation, we reasoned the missingness was likely due to the dominant signal from the coisolated ion species. We demonstrated the improvement by priming with two fragmentation modes for typical tryptic peptides, however, we anticipate that customizing fragmentation could be even more advantageous with other fragmentation strategies (e.g., ETD,³⁶ EThcD^38,39) for highly heterogeneous types of targets (e.g., HLA peptides,²³ phosphopeptides,²⁴ glycopeptides⁴¹). In addition, the latest advancement in machine learning enables the prediction of MS2 spectra across different collision energies and fragmentation modes,^56,57 which we believe will become great resources to guide the priming experiment of a Tomahto-based targeted proteomics assay.

Figure 5.

Tomahto assay with customized fragmentation further improves data completeness. Triplicate 60 min Tomahto analyses were performed with the 290 target peptides using the unprimed and primed fragmentation settings, respectively. (A) number of quantified peptides. Bars represent mean ± SD. (B) Venn diagram showing the overlap of quantified peptides between 60 min single-shot Tomahto analyses and the previously reported deep fractionated data set⁵⁵ obtained by analyzing 12 fractions and 1,080 min (18 h) total experiment time. Tomahto assay produced better coverage on the targeted peptides compared to the shotgun data set. (C) Peptide FYDSWESTVK was quantified only using the primed Tomahto setting. Its fragmentation pattern differed substantially between the default CID (NCE = 35%) and primed HCD (NCE = 27%). (D) The full scan events to quantify peptide FYDSWESTVK include a trigger MS2 scan on the STP, a target MS2 scan on the multiplexed endogenous counterpart, and an SPS-MS3 scan.

CONCLUSIONS

We present the upgraded Tomahto platform for real-time targeted proteomics, which integrates the FAIMS device for gas-phase fractionation and the latest sample multiplexing reagents, TMTpro18 and shTMTpro. Additionally, we have developed a software package, TomahtoPrimer, which determines the optimal fragmentation strategy in a peptide-specific manner. We have demonstrated that Tomahto can determine the FAIMS CV value for each target peptide in real-time and enhance the sensitivity and quantitative accuracy by effectively removing interfering signals. This improvement would be valuable, particularly when target peptides are of low abundance. Moreover, we highlighted that priming the fragmentation setting for each target significantly improves the success rate of a Tomahto assay. We envision the flexibility and customizability offered by Tomahto can be greatly beneficial for research involving nontryptic peptides (e.g., HLA) and those peptides bearing labile post-translational modifications.

Data Availability Statement

The raw files have been deposited at the MassIVE repository (https://massive.ucsd.edu/ProteoSAFe/static/massive.jsp) with the data set identifier MSV000095604.