Clinical biomarker discovery by SWATH-MS based label-free quantitative proteomics: impact of criteria for identification of differentiators and data normalization method.

BACKGROUND: SWATH-MS has emerged as the strategy of choice for biomarker discovery due to the proteome coverage achieved in acquisition and provision to re-interrogate the data. However, in quantitative analysis using SWATH, each sample from the comparison group is run individually in mass spectrometer and the resulting inter-run variation may influence relative quantification and identification of biomarkers. Normalization of data to diminish this variation thereby becomes an essential step in SWATH data processing. In most reported studies, data normalization methods used are those provided in instrument-based data analysis software or those used for microarray data. This study, for the first time provides an experimental evidence for selection of normalization method optimal for biomarker identification. METHODS: The efficiency of 12 normalization methods to normalize SWATH-MS data was evaluated based on statistical criteria in 'Normalyzer'-a tool which provides comparative evaluation of normalization by different methods. Further, the suitability of normalized data for biomarker discovery was assessed by evaluating the clustering efficiency of differentiators, identified from the normalized data based on p-value, fold change and both, by hierarchical clustering in Genesis software v.1.8.1. RESULTS: Conventional statistical criteria identified VSN-G as the optimal method for normalization of SWATH data. However, differentiators identified from VSN-G normalized data failed to segregate test and control groups. We thus assessed data normalized by eleven other methods for their ability to yield differentiators which segregate the study groups. Datasets in our study demonstrated that differentiators identified based on p-value from data normalized with Loess-R stratified the study groups optimally. CONCLUSION: This is the first report of experimentally tested strategy for SWATH-MS data processing with an emphasis on identification of clinically relevant biomarkers. Normalization of SWATH-MS data by Loess-R method and identification of differentiators based on p-value were found to be optimal for biomarker discovery in this study. The study also demonstrates the need to base the choice of normalization method on the application of the data.

Standardization approaches in absolute quantitative proteomics with mass spectrometry.

Mass spectrometry-based approaches have enabled important breakthroughs in quantitative proteomics in the last decades. This development is reflected in the better quantitative assessment of protein levels as well as to understand post-translational modifications and protein complexes and networks. Nowadays, the focus of quantitative proteomics shifted from the relative determination of proteins (ie, differential expression between two or more cellular states) to absolute quantity determination, required for a more-thorough characterization of biological models and comprehension of the proteome dynamism, as well as for the search and validation of novel protein biomarkers. However, the physico-chemical environment of the analyte species affects strongly the ionization efficiency in most mass spectrometry (MS) types, which thereby require the use of specially designed standardization approaches to provide absolute quantifications. Most common of such approaches nowadays include (i) the use of stable isotope-labeled peptide standards, isotopologues to the target proteotypic peptides expected after tryptic digestion of the target protein; (ii) use of stable isotope-labeled protein standards to compensate for sample preparation, sample loss, and proteolysis steps; (iii) isobaric reagents, which after fragmentation in the MS/MS analysis provide a final detectable mass shift, can be used to tag both analyte and standard samples; (iv) label-free approaches in which the absolute quantitative data are not obtained through the use of any kind of labeling, but from computational normalization of the raw data and adequate standards; (v) elemental mass spectrometry-based workflows able to provide directly absolute quantification of peptides/proteins that contain an ICP-detectable element. A critical insight from the Analytical Chemistry perspective of the different standardization approaches and their combinations used so far for absolute quantitative MS-based (molecular and elemental) proteomics is provided in this review.

Evaluation of NCI-7 Cell Line Panel as a Reference Material for Clinical Proteomics.

Reference materials are vital to benchmarking the reproducibility of clinical tests and essential for monitoring laboratory performance for clinical proteomics. The reference material utilized for mass spectrometric analysis of the human proteome would ideally contain enough proteins to be suitably representative of the human proteome, as well as exhibit a stable protein composition in different batches of sample regeneration. Previously, The Clinical Proteomic Tumor Analysis Consortium (CPTAC) utilized a PDX-derived comparative reference (CompRef) materials for the longitudinal assessment of proteomic performance; however, inherent drawbacks of PDX-derived material, including extended time needed to grow tumors and high level of expertise needed, have resulted in efforts to identify a new source of CompRef material. In this study, we examined the utility of using a panel of seven cancer cell lines, NCI-7 Cell Line Panel, as a reference material for mass spectrometric analysis of human proteome. Our results showed that not only is the NCI-7 material suitable for benchmarking laboratory sample preparation methods, but also NCI-7 sample generation is highly reproducible at both the global and phosphoprotein levels. In addition, the predicted genomic and experimental coverage of the NCI-7 proteome suggests the NCI-7 material may also have applications as a universal standard proteomic reference.

Calibration Using a Single-Point External Reference Material Harmonizes Quantitative Mass Spectrometry Proteomics Data between Platforms and Laboratories.

Mass spectrometry (MS) measurements are not inherently calibrated. Researchers use various calibration methods to assign meaning to arbitrary signal intensities and improve precision. Internal calibration (IC) methods use internal standards (IS) such as synthesized or recombinant proteins or peptides to calibrate MS measurements by comparing endogenous analyte signal to the signal from known IS concentrations spiked into the same sample. However, recent work suggests that using IS as IC introduces quantitative biases that affect comparison across studies because of the inability of IS to capture all sources of variation present throughout an MS workflow. Here, we describe a single-point external calibration strategy to calibrate signal intensity measurements to a common reference material, placing MS measurements on the same scale and harmonizing signal intensities between instruments, acquisition methods, and sites. We demonstrate data harmonization between laboratories and methodologies using this generalizable approach.

Human Prestin: A Candidate PE1 Protein Lacking Stringent Mass Spectrometric Evidence?

The evidence that any protein exists in the Human Proteome Project (HPP; protein evidence 1 or PE1) has revolved primarily (although not exclusively) around mass spectrometry (MS) (93% of PE1 proteins have MS evidence in the latest neXtProt release), with robust and stringent, well-curated metrics that have served the community well. This has led to a significant number of proteins still considered "missing" (i.e., PE2-4). Many PE2-4 proteins have MS evidence of unacceptable quality (small or not enough unitypic peptides and unacceptably high protein/peptide FDRs), transcriptomic, or antibody evidence. Here we use a Chromosome 7 PE2 example called Prestin to demonstrate that clear and robust criteria/metrics need to be developed for proteins that may not or cannot produce clear-cut MS evidence while possessing significant non-MS evidence, including disease-association data. Many of the PE2-4 proteins are inaccessible, spatiotemporally expressed in a limited way, or expressed at such a very low copy number as to be unable to be detected by current MS methodologies. We propose that the HPP community consider and lead a communal initiative to accelerate the discovery and characterization of these types of "missing" proteins.

Evaluating the Characteristics of Reporter Ion Signal Acquired in the Orbitrap Analyzer for Isobaric Mass Tag Proteome Quantification Experiments.

Multiplexed quantification with isobaric chemical tags (e.g., TMT, iTRAQ) provides a robust and efficient means to comparatively examine proteome dynamics between several biological states using a mass spectrometer (MS). The quantitative nature of isobaric tags necessitates strict validation of the observed ion signals in the chosen MS detector before differential patterns are extracted between biological states. We present an in-depth analysis of isobaric tag data acquired on current generation Orbitrap MS hardware to illustrate pitfalls in acquisition settings that can negatively impact results. We establish, for the first time, the presence of a notch, a region of no observed values, in the reporter ion distributions from isobaric-labeled peptide mixtures acquired on these instruments. We determine that this notch is present in published data across a wide range of instruments of the same or different type and is isolated to the Orbitrap mass analyzer. We demonstrate that the impact of the notch can be minimized using manipulations of Orbitrap scan parameters and on-column injection amounts. Lastly, using a mixture of synthetic standard peptides we investigated the impact on identification rates and quantification precision. Together, these data highlight an important phenomenon that negatively impacts peptide identification and quantification in the Orbitrap analyzer as well as outlining guidelines to follow to ensure minimization of MS-induced artifacts in isobaric tag experiments resulting from the notch.

Correlation Patterns in Experimental Data Are Affected by Normalization Procedures: Consequences for Data Analysis and Network Inference.

Normalization is a fundamental step in data processing to account for the sample-to-sample variation observed in biological samples. However, data structure is affected by normalization. In this paper, we show how, and to what extent, the correlation structure is affected by the application of 11 different normalization procedures. We also discuss the consequences for data analysis and interpretation, including principal component analysis, partial least-squares discrimination, and the inference of metabolite-metabolite association networks.

ABRF Proteome Informatics Research Group (iPRG) 2015 Study: Detection of Differentially Abundant Proteins in Label-Free Quantitative LC-MS/MS Experiments.

Detection of differentially abundant proteins in label-free quantitative shotgun liquid chromatography-tandem mass spectrometry (LC-MS/MS) experiments requires a series of computational steps that identify and quantify LC-MS features. It also requires statistical analyses that distinguish systematic changes in abundance between conditions from artifacts of biological and technical variation. The 2015 study of the Proteome Informatics Research Group (iPRG) of the Association of Biomolecular Resource Facilities (ABRF) aimed to evaluate the effects of the statistical analysis on the accuracy of the results. The study used LC-tandem mass spectra acquired from a controlled mixture, and made the data available to anonymous volunteer participants. The participants used methods of their choice to detect differentially abundant proteins, estimate the associated fold changes, and characterize the uncertainty of the results. The study found that multiple strategies (including the use of spectral counts versus peak intensities, and various software tools) could lead to accurate results, and that the performance was primarily determined by the analysts' expertise. This manuscript summarizes the outcome of the study, and provides representative examples of good computational and statistical practice. The data set generated as part of this study is publicly available.

The Human Proteome Organization-Proteomics Standards Initiative Quality Control Working Group: Making Quality Control More Accessible for Biological Mass Spectrometry.

To have confidence in results acquired during biological mass spectrometry experiments, a systematic approach to quality control is of vital importance. Nonetheless, until now, only scattered initiatives have been undertaken to this end, and these individual efforts have often not been complementary. To address this issue, the Human Proteome Organization-Proteomics Standards Initiative has established a new working group on quality control at its meeting in the spring of 2016. The goal of this working group is to provide a unifying framework for quality control data. The initial focus will be on providing a community-driven standardized file format for quality control. For this purpose, the previously proposed qcML format will be adapted to support a variety of use cases for both proteomics and metabolomics applications, and it will be established as an official PSI format. An important consideration is to avoid enforcing restrictive requirements on quality control but instead provide the basic technical necessities required to support extensive quality control for any type of mass spectrometry-based workflow. We want to emphasize that this is an open community effort, and we seek participation from all scientists with an interest in this field.

Evaluation and Improvement of Quantification Accuracy in Isobaric Mass Tag-Based Protein Quantification Experiments.

The multiplexing capabilities of isobaric mass tag-based protein quantification, such as Tandem Mass Tags or Isobaric Tag for Relative and Absolute Quantitation have dramatically increased the scope of mass spectrometry-based proteomics studies. Not only does the technology allow for the simultaneous quantification of multiple samples in a single MS injection, but its seamless compatibility with extensive sample prefractionation methods allows for comprehensive studies of complex proteomes. However, reporter ion-based quantification has often been criticized for limited quantification accuracy due to interference from coeluting peptides and peptide fragments. In this study, we investigate the extent of this problem and propose an effective and easy-to-implement remedy that relies on spiking a 6-protein calibration mixture to the samples. We evaluated our ratio adjustment approach using two large scale TMT 10-plex data sets derived from a human cancer and noncancer cell line as well as E. coli cells grown at two different conditions. Furthermore, we analyzed a complex 2-proteome artificial sample mixture and investigated the precision of TMT and precursor ion intensity-based label free quantification. Studying the protein set identified by both methods, we found that differentially abundant proteins were assigned dramatically higher statistical significance when quantified using TMT. Data are available via ProteomeXchange with identifier PXD003346.

Automated SDS Depletion for Mass Spectrometry of Intact Membrane Proteins though Transmembrane Electrophoresis.

Membrane proteins are underrepresented in proteome analysis platforms because of their hydrophobic character, contributing to decreased solubility. Sodium dodecyl sulfate is a favored denaturant in proteomic workflows, facilitating cell lysis and protein dissolution; however, SDS impedes MS detection and therefore must be removed prior to analysis. Although strategies exist for SDS removal, they provide low recovery, purity, or reproducibility. Here we present a simple automated device, termed transmembrane electrophoresis (TME), incorporating the principles of membrane filtration, but with an applied electric current to ensure near-complete (99.9%) removal of the surfactant, including protein-bound SDS. Intact proteins are recovered in solution phase in high yield (90-100%) within 1 h of operation. The strategy is applied to protein standards and proteome mixtures, including an enriched membrane fraction from E. coli, resulting in quality MS spectra free of SDS adducts. The TME platform is applicable to both bottom-up MS/MS as well as LC-ESI-MS analysis of intact proteins. SDS-depleted fractions reveal a similar number of protein identifications (285) compared wit a non-SDS control (280), being highly correlated in terms of protein spectral counts. This fully automated approach to SDS removal presents a viable tool for proteome sample processing ahead of MS analysis. Data are available via ProteomeXchange, identifier PXD003941.

Representation of selected-reaction monitoring data in the mzQuantML data standard.

The mzQuantML data standard was designed to capture the output of quantitative software in proteomics, to support submissions to public repositories, development of visualization software and pipeline/modular approaches. The standard is designed around a common core that can be extended to support particular types of technique through the release of semantic rules that are checked by validation software. The first release of mzQuantML supported four quantitative proteomics techniques via four sets of semantic rules: (i) intensity-based (MS(1) ) label free, (ii) MS(1) label-based (such as SILAC or N(15) ), (iii) MS(2) tag-based (iTRAQ or tandem mass tags), and (iv) spectral counting. We present an update to mzQuantML for supporting SRM techniques. The update includes representing the quantitative measurements, and associated meta-data, for SRM transitions, the mechanism for inferring peptide-level or protein-level quantitative values, and support for both label-based or label-free SRM protocols, through the creation of semantic rules and controlled vocabulary terms. We have updated the specification document for mzQuantML (version 1.0.1) and the mzQuantML validator to ensure that consistent files are produced by different exporters. We also report the capabilities for production of mzQuantML files from popular SRM software packages, such as Skyline and Anubis.

A proteome quality index.

We present the Proteome Quality Index (PQI; http://pqi-list.org), a much-needed resource for users of bacterial and eukaryotic proteomes. Completely sequenced genomes for which there is an available set of protein sequences (the proteome) are given a one- to five-star rating supported by 11 different metrics of quality. The database indexes over 3000 proteomes at the time of writing and is provided via a website for browsing, filtering and downloading. Previous to this work, there was no systematic way to account for the large variability in quality of the thousands of proteomes, and this is likely to have profoundly influenced the outcome of many published studies, in particular large-scale comparative analyses. The lack of a measure of proteome quality is likely due to the difficulty in producing one, a problem that we have approached by integrating multiple metrics. The continued development and improvement of the index will require the contribution of additional metrics by us and by others; the PQI provides a useful point of reference for the scientific community, but it is only the first step towards a 'standard' for the field.

ICan: an optimized ion-current-based quantification procedure with enhanced quantitative accuracy and sensitivity in biomarker discovery.

The rapidly expanding availability of high-resolution mass spectrometry has substantially enhanced the ion-current-based relative quantification techniques. Despite the increasing interest in ion-current-based methods, quantitative sensitivity, accuracy, and false discovery rate remain the major concerns; consequently, comprehensive evaluation and development in these regards are urgently needed. Here we describe an integrated, new procedure for data normalization and protein ratio estimation, termed ICan, for improved ion-current-based analysis of data generated by high-resolution mass spectrometry (MS). ICan achieved significantly better accuracy and precision, and lower false-positive rate for discovering altered proteins, over current popular pipelines. A spiked-in experiment was used to evaluate the performance of ICan to detect small changes. In this study E. coli extracts were spiked with moderate-abundance proteins from human plasma (MAP, enriched by IgY14-SuperMix procedure) at two different levels to set a small change of 1.5-fold. Forty-five (92%, with an average ratio of 1.71 ± 0.13) of 49 identified MAP protein (i.e., the true positives) and none of the reference proteins (1.0-fold) were determined as significantly altered proteins, with cutoff thresholds of ≥ 1.3-fold change and p ≤ 0.05. This is the first study to evaluate and prove competitive performance of the ion-current-based approach for assigning significance to proteins with small changes. By comparison, other methods showed remarkably inferior performance. ICan can be broadly applicable to reliable and sensitive proteomic survey of multiple biological samples with the use of high-resolution MS. Moreover, many key features evaluated and optimized here such as normalization, protein ratio determination, and statistical analyses are also valuable for data analysis by isotope-labeling methods.

Letter to the editor: SeqXML and OrthoXML: standards for sequence and orthology information.

There is a great need for standards in the orthology field. Users must contend with different ortholog data representations from each provider, and the providers themselves must independently gather and parse the input sequence data. These burdensome and redundant procedures make data comparison and integration difficult. We have designed two XML-based formats, SeqXML and OrthoXML, to solve these problems. SeqXML is a lightweight format for sequence records-the input for orthology prediction. It stores the same sequence and metadata as typical FASTA format records, but overcomes common problems such as unstructured metadata in the header and erroneous sequence content. XML provides validation to prevent data integrity problems that are frequent in FASTA files. The range of applications for SeqXML is broad and not limited to ortholog prediction. We provide read/write functions for BioJava, BioPerl, and Biopython. OrthoXML was designed to represent ortholog assignments from any source in a consistent and structured way, yet cater to specific needs such as scoring schemes or meta-information. A unified format is particularly valuable for ortholog consumers that want to integrate data from numerous resources, e.g. for gene annotation projects. Reference proteomes for 61 organisms are already available in SeqXML, and 10 orthology databases have signed on to OrthoXML. Adoption by the entire field would substantially facilitate exchange and quality control of sequence and orthology information.

18O-labeled proteome reference as global internal standards for targeted quantification by selected reaction monitoring-mass spectrometry.

Selected reaction monitoring (SRM)-MS is an emerging technology for high throughput targeted protein quantification and verification in biomarker discovery studies; however, the cost associated with the application of stable isotope-labeled synthetic peptides as internal standards can be prohibitive for screening a large number of candidate proteins as often required in the preverification phase of discovery studies. Herein we present a proof of concept study using an (18)O-labeled proteome reference as global internal standards (GIS) for SRM-based relative quantification. The (18)O-labeled proteome reference (or GIS) can be readily prepared and contains a heavy isotope ((18)O)-labeled internal standard for every possible tryptic peptide. Our results showed that the percentage of heavy isotope ((18)O) incorporation applying an improved protocol was >99.5% for most peptides investigated. The accuracy, reproducibility, and linear dynamic range of quantification were further assessed based on known ratios of standard proteins spiked into the labeled mouse plasma reference. Reliable quantification was observed with high reproducibility (i.e. coefficient of variance <10%) for analyte concentrations that were set at 100-fold higher or lower than those of the GIS based on the light ((16)O)/heavy ((18)O) peak area ratios. The utility of (18)O-labeled GIS was further illustrated by accurate relative quantification of 45 major human plasma proteins. Moreover, quantification of the concentrations of C-reactive protein and prostate-specific antigen was illustrated by coupling the GIS with standard additions of purified protein standards. Collectively, our results demonstrated that the use of (18)O-labeled proteome reference as GIS provides a convenient, low cost, and effective strategy for relative quantification of a large number of candidate proteins in biological or clinical samples using SRM.

Ten years of standardizing proteomic data: a report on the HUPO-PSI Spring Workshop: April 12-14th, 2012, San Diego, USA.

The Human Proteome Organisation Proteomics Standards Initiative (HUPO-PSI) was established in 2002 with the aim of defining community standards for data representation in proteomics and facilitating data comparison, exchange and verification. Over the last 10 years significant advances have been made, with common data standards now published and implemented in the field of both mass spectrometry and molecular interactions. The 2012 meeting further advanced this work, with the mass spectrometry groups finalising approaches to capturing the output from recent developments in the field, such as quantitative proteomics and SRM. The molecular interaction group focused on improving the integration of data from multiple resources. Both groups united with a guest work track, organized by the HUPO Technology/Standards Committee, to formulate proposals for data submissions from the HUPO Human Proteome Project and to start an initiative to collect standard experimental protocols.

Deterministic protein inference for shotgun proteomics data provides new insights into Arabidopsis pollen development and function.

Pollen, the male gametophyte of flowering plants, represents an ideal biological system to study developmental processes, such as cell polarity, tip growth, and morphogenesis. Upon hydration, the metabolically quiescent pollen rapidly switches to an active state, exhibiting extremely fast growth. This rapid switch requires relevant proteins to be stored in the mature pollen, where they have to retain functionality in a desiccated environment. Using a shotgun proteomics approach, we unambiguously identified approximately 3500 proteins in Arabidopsis pollen, including 537 proteins that were not identified in genetic or transcriptomic studies. To generate this comprehensive reference data set, which extends the previously reported pollen proteome by a factor of 13, we developed a novel deterministic peptide classification scheme for protein inference. This generally applicable approach considers the gene model-protein sequence-protein accession relationships. It allowed us to classify and eliminate ambiguities inherently associated with any shotgun proteomics data set, to report a conservative list of protein identifications, and to seamlessly integrate data from previous transcriptomics studies. Manual validation of proteins unambiguously identified by a single, information-rich peptide enabled us to significantly reduce the false discovery rate, while keeping valuable identifications of shorter and lower abundant proteins. Bioinformatic analyses revealed a higher stability of pollen proteins compared to those of other tissues and implied a protein family of previously unknown function in vesicle trafficking. Interestingly, the pollen proteome is most similar to that of seeds, indicating physiological similarities between these developmentally distinct tissues.

Operator- and software-related post-experimental variability and source of error in 2-DE analysis.

In the field of proteomics, several approaches have been developed for separating proteins and analyzing their differential relative abundance. One of the oldest, yet still widely used, is 2-DE. Despite the continuous advance of new methods, which are less demanding from a technical standpoint, 2-DE is still compelling and has a lot of potential for improvement. The overall variability which affects 2-DE includes biological, experimental, and post-experimental (software-related) variance. It is important to highlight how much of the total variability of this technique is due to post-experimental variability, which, so far, has been largely neglected. In this short review, we have focused on this topic and explained that post-experimental variability and source of error can be further divided into those which are software-dependent and those which are operator-dependent. We discuss these issues in detail, offering suggestions for reducing errors that may affect the quality of results, summarizing the advantages and drawbacks of each approach.