Reinspection of a Clinical Proteomics Tumor Analysis Consortium (CPTAC) Dataset with Cloud Computing Reveals Abundant Post-Translational Modifications and Protein Sequence Variants.

The Clinical Proteomic Tumor Analysis Consortium (CPTAC) has provided some of the most in-depth analyses of the phenotypes of human tumors ever constructed. Today, the majority of proteomic data analysis is still performed using software housed on desktop computers which limits the number of sequence variants and post-translational modifications that can be considered. The original CPTAC studies limited the search for PTMs to only samples that were chemically enriched for those modified peptides. Similarly, the only sequence variants considered were those with strong evidence at the exon or transcript level. In this multi-institutional collaborative reanalysis, we utilized unbiased protein databases containing millions of human sequence variants in conjunction with hundreds of common post-translational modifications. Using these tools, we identified tens of thousands of high-confidence PTMs and sequence variants. We identified 4132 phosphorylated peptides in nonenriched samples, 93% of which were confirmed in the samples which were chemically enriched for phosphopeptides. In addition, our results also cover 90% of the high-confidence variants reported by the original proteogenomics study, without the need for sample specific next-generation sequencing. Finally, we report fivefold more somatic and germline variants that have an independent evidence at the peptide level, including mutations in ERRB2 and BCAS1. In this reanalysis of CPTAC proteomic data with cloud computing, we present an openly available and searchable web resource of the highest-coverage proteomic profiling of human tumors described to date.

Cloud Computing Based Immunopeptidomics Utilizing Community Curated Variant Libraries Simplifies and Improves Neo-Antigen Discovery in Metastatic Melanoma.

Unique peptide neo-antigens presented on the cell surface are attractive targets for researchers in nearly all areas of personalized medicine. Cells presenting peptides with mutated or other non-canonical sequences can be utilized for both targeted therapies and diagnostics. Today's state-of-the-art pipelines utilize complementary proteogenomic approaches where RNA or ribosomal sequencing data helps to create libraries from which tandem mass spectrometry data can be compared. In this study, we present an alternative approach whereby cloud computing is utilized to power neo-antigen searches against community curated databases containing more than 7 million human sequence variants. Using these expansive databases of high-quality sequences as a reference, we reanalyze the original data from two previously reported studies to identify neo-antigen targets in metastatic melanoma. Using our approach, we identify 79 percent of the non-canonical peptides reported by previous genomic analyses of these files. Furthermore, we report 18-fold more non-canonical peptides than previously reported. The novel neo-antigens we report herein can be corroborated by secondary analyses such as high predicted binding affinity, when analyzed by well-established tools such as NetMHC. Finally, we report 738 non-canonical peptides shared by at least five patient samples, and 3258 shared across the two studies. This illustrates the depth of data that is present, but typically missed by lower statistical power proteogenomic approaches. This large list of shared peptides across the two studies, their annotation, non-canonical origin, as well as MS/MS spectra from the two studies are made available on a web portal for community analysis.

The utilization of the search engine, Bolt, to decrease search time and increase peptide identifications in hydroxyl radical protein footprinting-based workflows.

Hydroxyl radical protein footprinting (HRPF) utilizes hydroxyl radicals to covalently modify solvent exposed regions of proteins. When coupled with mass spectrometry, HRPF can provide insightful information on protein structural changes including inside cells. However, the complex mixture of proteins and modifications makes identification a complicated task. To search all of the HRPF-induced modification combinations across the full proteome, requires substantial computational power and still can take days to search. To drastically decrease processing time and improve identifications, a novel cloud-based search engine, Bolt, was used to search for HRPF modifications in comparison to a commonly used search engine, Sequest. A 35% increase in the identification of modified peptides was observed in Bolt compared to Sequest with a decrease in computation time.

Digging deeper into the immunopeptidome: characterization of post-translationally modified peptides presented by MHC I.

Peptides presented by MHC molecules on the cell surface, or the immunopeptidome, play an important role in the adaptive arm of the immune response. Antigen processing for MHC class I molecules is a ubiquitous pathway present in all nucleated cells which generates and presents peptides of both self and non-self-origin. Peptides with post-translational modifications represent one category of peptides presented by MHC class I molecules. However, owing to the complexity of self-peptides presented by cells, the diversity of peptides with post-translational modifications is not well-studied. In this study, we carried out MHC Class I immunopeptidomics analysis of Loucy T-cell leukemia and A375 malignant melanoma cell line to characterize the diversity of post-translational modifications of MHC class I-bound peptides. Using high resolution mass spectrometry, we identified 25,761 MHC-bound peptides across both cell lines using Bolt and Sequest search engines. The enrichment method was highly specific as ~ 90% of the peptides were of typical length (8-12 amino acids long) and the motifs were expected based on previously reported motifs for MHC I alleles. Among the MHC-bound peptides, we identified phosphorylation as a major post-translational modification followed by deamidation. We observed site-specific localization of these post-translational modifications, at position P4 for phosphorylated peptides and position P3 for deamidated peptides. We identified a smaller number of peptides with acetylated and methylated lysine, possibly due to very low stoichiometric levels of these PTMs compared to phosphorylation and deamidation. Using PEAKS de novo sequencing algorithm, we identified spliced peptides that accounted for ~ 5-7% of MHC-bound peptides that were otherwise similar in their features as normal MHC-bound peptides. We validated the identity of several post-translationally modified peptides and spliced peptides through mass spectrometric analysis of synthetic peptides. Our study confirms post-translationally modified peptides to be present at low stoichiometric levels along with unusual spliced peptides through unbiased identification using high resolution mass spectrometry. Supplementary Information: The online version contains supplementary material available at 10.1007/s42485-021-00066-x.

Quantitative assessment of successive carbohydrate additions to the clustered O-glycosylation sites of IgA1 by glycosyltransferases.

Mucin-type O-glycosylation occurs on many proteins that transit the Golgi apparatus. These glycans impact structure and function of many proteins and have important roles in cellular biosynthetic processes, signaling and differentiation. Although recent technological advances have enhanced our ability to profile glycosylation of glycoproteins, limitations in the understanding of the biosynthesis of these glycan structures remain. Some of these limitations stem from the difficulty to track the biosynthetic process of mucin-type O-glycosylation, especially when glycans occur in dense clusters in repeat regions of proteins, such as the mucins or immunoglobulin A1 (IgA1). Here, we describe a series of nano-liquid chromatography (LC)-mass spectrometry (MS) analyses that demonstrate the range of glycosyltransferase enzymatic activities involved in the biosynthesis of clustered O-glycans on IgA1. By utilizing nano-LC-MS relative quantitation of in vitro reaction products, our results provide unique insights into the biosynthesis of clustered IgA1 O-glycans. We have developed a workflow to determine glycoform-specific apparent rates of a human UDP-N-acetylgalactosamine:polypeptide N-acetylgalactosaminyltrasnfersase (GalNAc-T EC 2.4.1.41) and demonstrated how pre-existing glycans affect subsequent activity of glycosyltransferases, such as core 1 galactosyltransferase and α2,3- and α2,6-specific sialyltransferases, in successive additions in the biosynthesis of clustered O-glycans. In the context of IgA1, these results have potential to provide insight into the molecular mechanisms implicated in the pathogenesis of IgA nephropathy, an autoimmune renal disease involving aberrant IgA1 O-glycosylation. In a broader sense, these methods and workflows are applicable to the studies of the concerted and competing functions of other glycosyltransferases that initiate and extend mucin-type core 1 clustered O-glycosylation.

Bolt: a New Age Peptide Search Engine for Comprehensive MS/MS Sequencing Through Vast Protein Databases in Minutes.

Recent increases in mass spectrometry speed, sensitivity, and resolution now permit comprehensive proteomics coverage. However, the results are often hindered by sub-optimal data processing pipelines. In almost all MS/MS peptide search engines, users must limit their search space to a canonical database due to time constraints and q value considerations, but this typically does not reflect the individual genetic variations of the organism being studied. In addition, engines will nearly always assume the presence of only fully tryptic peptides and limit PTMs to a handful. Even on high-performance servers, these search engines are computationally expensive, and most users decide to dial back their search parameters. We present Bolt, a new cloud-based search engine that can search more than 900,000 protein sequences (canonical, isoform, mutations, and contaminants) with 41 post-translation modifications and N-terminal and C-terminal partial tryptic search in minutes on a standard configuration laptop. Along with increases in speed, Bolt provides an additional benefit of improvement in high-confidence identifications. Sixty-one percent of peptides uniquely identified by Bolt may be validated by strong fragmentation patterns, compared with 13% of peptides uniquely identified by SEQUEST and 6% of peptides uniquely identified by Mascot. Furthermore, 30% of unique Bolt identifications were verified by all three software on the longer gradient analysis, compared with only 20% and 27% for SEQUEST and Mascot identifications respectively. Bolt represents, to the best of our knowledge, the first fully scalable, cloud-based quantitative proteomic solution that can be operated within a user-friendly GUI interface. Data are available via ProteomeXchange with identifier PXD012700.

Detection and verification of 2.3 million cancer mutations in NCI60 cancer cell lines with a cloud search engine.

Today we have unprecedented access to human genomic and proteomic data that appear to be rapidly approaching our current understanding of comprehensive coverage. Combining genomic information with shotgun proteomics remains challenging due to the large increase in proteomics search space. However, making this connection between genomic and proteomic information is critical for cancer studies to vaccine development. Furthermore, as we progress towards personalized medicine, it will be essential for proteomics analysis to identify individual mutations and variants in order to fully understand protein networks and to develop personalized therapies. While these advantages are well-established, only a few studies have demonstrated the successful integration of proteomic data with large genomic input. We present and examine the abilities of Bolt, a new cloud-based proteomics search engine to search for the presence of over 2.3 million known cancer mutations in a matter of minutes while still performing a standard proteomics search that includes 31 post translational modifications. We use previously published proteomics data sets and identify mutations that are verified using genomic studies as well as previous proteomics efforts. Our results also emphasize the need to search for mutations in a comprehensive manner while still searching for both common and rare PTMs. SIGNIFICANCE: We present and examine the abilities of Bolt, a new cloud-based proteomics search engine to search for the presence of over 2.3 million known cancer mutations in a matter of minutes while still performing a standard proteomics search that includes 31 post translational modifications. No other proteomics search software can do so.

Biopsy Based Proteomic Assay Predicts Risk of Biochemical Recurrence after Radical Prostatectomy.

PURPOSE: Current clinicopathological parameters are insufficient to predict the likelihood of biochemical recurrence in patients with prostate cancer after radical prostatectomy. Such information may help identify patients who would likely benefit from adjuvant radiotherapy rather than active surveillance. A multiplex proteomic assay, previously tested on biopsies and found to be predictive of favorable or unfavorable pathology at radical prostatectomy, was assessed for its predictive value to identify patients at higher risk for biochemical relapse. MATERIALS AND METHODS: Proteomic assays from core needle biopsies of 288 men who subsequently underwent radical prostatectomy at CHUM (Centre hospitalier de l'Université de Montréal) were evaluated for the prediction of subsequent biochemical recurrence. RESULTS: Of the 288 men, biochemical relapse was observed in 47 (16.3%) and metastases were found in 5 (1.7%). Median followup was 68.5 months. The proteomic assay clearly separated patients into 3 categories, including those at low, intermediate and high risk for biochemical relapse (p = 0.0007). Assay scores predicted biochemical relapse on univariate analysis (HR 1.724, p = 0.0002 per 20% change in score), significantly better than other preoperative prognostic parameters. Additionally, the assay score had a significantly higher p value when combined with clinical National Comprehensive Cancer Network® stage compared to stage alone (HR 1.579, p = 0.0017 per 20% change in score). CONCLUSIONS: A protein based assay score derived from diagnostic needle biopsy has strong predictive ability for biochemical relapse after surgery. These results suggest that this assay score can be used at the diagnostic stage to identify patients in whom prostate cancer is potentially more biologically aggressive and active treatment should be considered.

Hybrid data acquisition and processing strategies with increased throughput and selectivity: pSMART analysis for global qualitative and quantitative analysis.

Data-dependent acquisition (DDA) and data-independent acquisition strategies (DIA) have both resulted in improved understanding of proteomics samples. Both strategies have advantages and disadvantages that are well-published, where DDA is typically applied for deep discovery and DIA may be used to create sample records. In this paper, we present a hybrid data acquisition and processing strategy (pSMART) that combines the strengths of both techniques and provides significant benefits for qualitative and quantitative peptide analysis. The performance of pSMART is compared to published DIA strategies in an experiment that allows the objective assessment of DIA performance with respect to interrogation of previously acquired MS data. The results of this experiment demonstrate that pSMART creates fewer decoy hits than a standard DIA strategy. Moreover, we show that pSMART is more selective, sensitive, and reproducible than either standard DIA or DDA strategies alone.

Rapid development of sensitive, high-throughput, quantitative and highly selective mass spectrometric targeted immunoassays for clinically important proteins in human plasma and serum.

OBJECTIVES: The aim of this study was to develop high-throughput, quantitative and highly selective mass spectrometric, targeted immunoassays for clinically important proteins in human plasma or serum. DESIGN AND METHODS: The described method coupled mass spectrometric immunoassay (MSIA), a previously developed technique for immunoenrichment on a monolithic microcolumn activated with an anti-protein antibody and fixed in a pipette tip, to selected reaction monitoring (SRM) detection and accurate quantification of targeted peptides, including clinically relevant sequence or truncated variants. RESULTS: In this report, we demonstrate the rapid development of MSIA-SRM assays for sixteen different target proteins spanning seven different clinically important areas (including neurological, Alzheimer's, cardiovascular, endocrine function, cancer and other diseases) and ranging in concentration from pg/mL to mg/mL. The reported MSIA-SRM assays demonstrated high sensitivity (within published clinical ranges), precision, robustness and high-throughput as well as specific detection of clinically relevant isoforms for many of the target proteins. Most of the assays were tested with bona-fide clinical samples. In addition, positive correlations, (R2 0.67-0.87, depending on the target peptide), were demonstrated for MSIA-SRM assay data with clinical analyzer measurements of parathyroid hormone (PTH) and insulin growth factor 1 (IGF1) in clinical sample cohorts. CONCLUSIONS: We have presented a practical and scalable method for rapid development and deployment of MS-based SRM assays for clinically relevant proteins and measured levels of the target analytes in bona fide clinical samples. The method permits the specific quantification of individual protein isoforms and addresses the difficult problem of protein heterogeneity in clinical proteomics applications.

Depletion of nuclear histone H2A variants is associated with chronic DNA damage signaling upon drug-evoked senescence of human somatic cells.

Cellular senescence is associated with global chromatin changes, altered gene expression, and activation of chronic DNA damage signaling. These events ultimately lead to morphological and physiological transformations in primary cells. In this study, we show that chronic DNA damage signals caused by genotoxic stress impact the expression of histones H2A family members and lead to their depletion in the nuclei of senescent human fibroblasts. Our data reinforce the hypothesis that progressive chromatin destabilization may lead to the loss of epigenetic information and impaired cellular function associated with chronic DNA damage upon drug-evoked senescence. We propose that changes in the histone biosynthesis and chromatin assembly may directly contribute to cellular aging. In addition, we also outline the method that allows for quantitative and unbiased measurement of these changes.

Interlaboratory reproducibility of selective reaction monitoring assays using multiple upfront analyte enrichment strategies.

Over the past few years, mass spectrometry has emerged as a technology to complement and potentially replace standard immunoassays in routine clinical core laboratories. Application of mass spectrometry to protein and peptide measurement can provide advantages including high sensitivity, the ability to multiplex analytes, and high specificity at the amino acid sequence level. In our previous study, we demonstrated excellent reproducibility of mass spectrometry-selective reaction monitoring (MS-SRM) assays when applying standardized standard operating procedures (SOPs) to measure synthetic peptides in a complex sample, as lack of reproducibility has been a frequent criticism leveled at the use of mass spectrometers in the clinical laboratory compared to immunoassays. Furthermore, an important caveat of SRM-based assays for proteins is that many low-abundance analytes require some type of enrichment before detection with MS. This adds a level of complexity to the procedure and the potential for irreproducibility increases, especially across different laboratories with different operators. The purpose of this study was to test the interlaboratory reproducibility of SRM assays with various upfront enrichment strategies and different types of clinical samples (representing real-world body fluids commonly encountered in routine clinical laboratories). Three different, previously published enrichment strategies for low-abundance analytes and a no-enrichment strategy for high-abundance analytes were tested across four different laboratories using different liquid chromatography-SRM (LC-SRM) platforms and previously developed SOPs. The results demonstrated that these assays were indeed reproducible with coefficients of variation of less than 30% for the measurement of important clinical proteins across all four laboratories in real world samples.

Evaluation of absolute peptide quantitation strategies using selected reaction monitoring.

The use of internal peptide standards in selected reaction monitoring experiments enables absolute quantitation. Here, we describe three approaches addressing calibration of peptide concentrations in complex matrices and assess their performance in terms of trueness and precision. The simplest approach described is single reference point quantitation where a heavy peptide is spiked into test samples and the endogenous analyte quantified relative to the heavy peptide internal standard. We refer to the second approach as normal curve quantitation. Here, a constant amount of heavy peptide and a varying amount of light peptide are spiked into matrix to construct a calibration curve. This accounts for matrix effects but due to the presence of endogenous analyte, it is usually not possible to determine the lower LOQ. We refer to the third method as reverse curve quantitation. Here, a constant amount of light peptide and a varying amount of heavy peptide are spiked into matrix to construct a calibration curve. Because there is no contribution to the heavy peptide signal from endogenous analyte, it is possible to measure the equivalent of a blank sample and determine LOQ. These approaches are applied to human plasma samples and used to assay peptides of a set of apolipoproteins.

Simultaneous analysis of glycosylated and sialylated prostate-specific antigen revealing differential distribution of glycosylated prostate-specific antigen isoforms in prostate cancer tissues.

Aberrant protein glycosylation has been shown to be associated with disease progression and can be potentially useful as a biomarker if disease-specific glycosylation can be identified. However, high-throughput quantitative analysis of protein glycosylation derived from clinical specimens presents technical challenges due to the typically high complexity of biological samples. In this study, a mass spectrometry-based analytical method was developed to measure different glycosylated forms of glycoproteins from complex biological samples by coupling glycopeptide extraction strategy for specific glycosylation with selected reaction monitoring (SRM). Using this method, we monitored glycosylated and sialylated prostate-specific antigen (PSA) in prostate cancer and noncancer tissues. Results of this study demonstrated that the relative abundance of glycosylated PSA isoforms were not correlated with total PSA protein levels measured in the same prostate cancer tissue samples by clinical immunoassay. Furthermore, the sialylated PSA was differentially distributed in cancer and noncancer tissues. These data suggest that differently glycosylated isoforms of glycoproteins can be quantitatively analyzed and may provide unique information for clinically relevant studies.

Increased selectivity, analytical precision, and throughput in targeted proteomics.

Proteomics is gradually complementing large shotgun qualitative studies with hypothesis-driven quantitative experiments. Targeted analyses performed on triple quadrupole instruments in selected reaction monitoring mode are characterized by a high degree of selectivity and low limit of detection; however, the concurrent analysis of multiple analytes occurs at the expense of sensitivity because of reduced dwell time and/or selectivity due to limitation to a few transitions. A new data acquisition paradigm is presented in which selected reaction monitoring is performed in two ways to simultaneously quantify and confirm the identity of the targeted peptides. A first set of primary transitions is continuously monitored during a predetermined elution time window to precisely quantify each peptide. In addition, a set of six to eight transitions is acquired in a data-dependent event, triggered when all the primary transitions exceed a preset threshold. These additional transitions are used to generate composite tandem mass spectra to formally confirm the identity of the targeted peptides. This technique was applied to analyze the tryptic digest of a yeast lysate to demonstrate the performance of the technique. We showed a limit of detection down to tens of attomoles injected and a throughput exceeding 6000 transitions in one 60-min experiment. The technique was integrated into a linear work flow, including experimental design, data acquisition, and data evaluation, enabling large scale proteomic studies.

Measurement of protein phosphorylation stoichiometry by selected reaction monitoring mass spectrometry.

The stoichiometry of protein phosphorylation at specific amino acid sites may be used to infer on the significance of the modification, and its biological function in the cell. However, detection and quantification of phosphorylation stoichiometry in tissue remain a significant challenge. Here we describe a strategy for highly sensitive, label-free quantification of protein phosphorylation stoichiometry. Method development included the analysis of synthetic peptides in order to determine constants to relate the mass spectrometry signals of cognate peptide/phosphopeptide pairs, and the detection of the cognate peptides by using high resolution Fourier Transform mass spectrometry (FTMS) and selected reaction monitoring mass spectrometry (SRM). By analyzing extracted ion currents by FTMS, the phosphorylation stoichiometries of two tyrosine residues (tyrosine-194 and tyrosine-397) in the protein tyrosine kinase Lyn were determined in transfected human HEK293T cells and two cultured human multiple myeloma strains. To achieve high sensitivity to measure phosphorylation stoichiometry in tissue, SRM methods were developed and applied for the analysis of phosphorylation stoichiometries of Lyn phospho-sites in multiple myeloma xenograft tumors. Western immuno-blotting was used to verify mass spectrometry findings. The SRM method has potential applications in analyzing clinical samples wherein protein phosphorylation stoichiometries may represent important pharmacodynamic biomarkers.

Quantitative phosphoproteomic analysis of the STAT3/IL-6/HIF1alpha signaling network: an initial study in GSC11 glioblastoma stem cells.

Initiation and maintenance of several cancers including glioblastoma (GBM) may be driven by a small subset of cells called cancer stem cells (CSCs). CSCs may provide a repository of cells in tumor cell populations that are refractory to chemotherapeutic agents developed for the treatment of tumors. STAT3 is a key transcription factor associated with regulation of multiple stem cell types. Recently, a novel autocrine loop (IL-6/STAT3/HIF1alpha) has been observed in multiple tumor types (pancreatic, prostate, lung, and colon). The objective of this study was to probe perturbations of this loop in a glioblastoma cancer stem cell line (GSC11) derived from a human tumor by use of a JAK2/STAT3 phosphorylation inhibitor (WP1193), IL-6 stimulation, and hypoxia. A quantitative phosphoproteomic approach that employed phosphoprotein enrichment, chemical tagging with isobaric tags, phosphopeptide enrichment, and tandem mass spectrometry in a high-resolution instrument was applied. A total of 3414 proteins were identified in this study. A rapid Western blotting technique (<1 h) was used to confirm alterations in key protein expression and phosphorylation levels observed in the mass spectrometric experiments. About 10% of the phosphoproteins were linked to the IL-6 pathway, and the majority of remaining proteins could be assigned to other interlinked networks. By multiple comparisons between the sample conditions, we observed expected changes and gained novel insights into the contribution of each factor to the IL6/STAT3/HIF1alpha autocrine loop and the CSC response to perturbations by hypoxia, inhibition of STAT3 phosphorylation, and IL-6 stimulation.

Epidermal growth factor receptor phosphorylation sites Ser991 and Tyr998 are implicated in the regulation of receptor endocytosis and phosphorylations at Ser1039 and Thr1041.

Aberrant expression, activation, and down-regulation of the epidermal growth factor receptor (EGFR) have causal roles in many human cancers, and post-translational modifications including phosphorylation and ubiquitination and protein-protein interactions directly modulate EGFR function. Quantitative mass spectrometric analyses including selected reaction monitoring (also known as multiple reaction monitoring) were applied to the EGFR and associated proteins. In response to epidermal growth factor (EGF) stimulation of cells, phosphorylations at EGFR Ser(991) and Tyr(998) accumulated more slowly than at receptor sites involved in RAS-ERK signaling. Phosphorylation-deficient mutant receptors S991A and Y998F activated ERK in response to EGF but were impaired for receptor endocytosis. Consistent with these results, the mutant receptors retained a network of interactions with known signaling proteins including EGF-stimulated binding to the adaptor GRB2. Compared with wild type EGFR the Y998F variant had diminished EGF-stimulated interaction with the ubiquitin E3 ligase CBL, and the S991A variant had decreased associated ubiquitin. The endocytosis-defective mutant receptors were found to have elevated phosphorylation at positions Ser(1039) and Thr(1041). These residues reside in a serine/threonine-rich region of the receptor previously implicated in p38 mitogen-activated protein kinase-dependent stress/cytokine-induced EGFR internalization and recycling (Zwang, Y., and Yarden, Y. (2006) p38 MAP kinase mediates stress-induced internalization of EGFR: implications for cancer chemotherapy. EMBO J. 25, 4195-4206). EGF-induced phosphorylations at Ser(1039) and Thr(1041) were blocked by treatment of cells with SB-202190, a selective inhibitor of p38. These results suggest that coordinated phosphorylation of EGFR involving sites Tyr(998), Ser(991), Ser(1039), and Thr(1041) governs the trafficking of EGF receptors. This reinforces the notion that EGFR function is manifest through spatially and temporally controlled protein-protein interactions and phosphorylations.

Signal maps for mass spectrometry-based comparative proteomics.

Mass spectrometry-based proteomic experiments, in combination with liquid chromatography-based separation, can be used to compare complex biological samples across multiple conditions. These comparisons are usually performed on the level of protein lists generated from individual experiments. Unfortunately given the current technologies, these lists typically cover only a small fraction of the total protein content, making global comparisons extremely limited. Recently approaches have been suggested that are built on the comparison of computationally built feature lists instead of protein identifications. Although these approaches promise to capture a bigger spectrum of the proteins present in a complex mixture, their success is strongly dependent on the correctness of the identified features and the aligned retention times of these features across multiple experiments. In this experimental-computational study, we went one step further and performed the comparisons directly on the signal level. First signal maps were constructed that associate the experimental signals across multiple experiments. Then a feature detection algorithm used this integrated information to identify those features that are discriminating or common across multiple experiments. At the core of our approach is a score function that faithfully recognizes mass spectra from similar peptide mixtures and an algorithm that produces an optimal alignment (time warping) of the liquid chromatography experiments on the basis of raw MS signal, making minimal assumptions on the underlying data. We provide experimental evidence that suggests uniqueness and correctness of the resulting signal maps even on low accuracy mass spectrometers. These maps can be used for a variety of proteomic analyses. Here we illustrate the use of signal maps for the discovery of diagnostic biomarkers. An imple-mentation of our algorithm is available on our Web server.