NewStrategies and Challenges in Lung Proteomics andMetabolomics An Official American Thoracic Society Workshop Report

This document presents the proceedings from the workshop entitled, “New Strategies and Challenges in Lung Proteomics and Metabolomics” held February 4th–5th, 2016, in Denver, Colorado. It was sponsored by the National Heart Lung Blood Institute, the American Thoracic Society, the Colorado Biological Mass Spectrometry Society, and National Jewish Health. The goal of this workshop was to convene, for the first time, relevant experts in lung proteomics and metabolomics to discuss and overcome specific challenges in these fields that are unique to the lung. The main objectives of this workshop were to identify, review, and/or understand: (1) emerging technologies in metabolomics and proteomics as applied to the study of the lung; (2) the unique composition and challenges of lung-specific biological specimens for metabolomic and proteomic analysis; (3) the diverse informatics approaches and databases unique to metabolomics and proteomics, with special emphasis on the lung; (4) integrative platforms across genetic and genomic databases that can be applied to lung-related metabolomic and proteomic studies; and (5) the clinical applications of proteomics and metabolomics. The major findings and conclusions of this workshop are summarized at the end of the report, and outline the progress and challenges that face these rapidly advancing fields.


Overview: The Role of Proteomics and Metabolomics in Systems Biology
Personalized disease risk and drug response predictions based on genomic sequences now represent a cornerstone of precision medicine, and have also been successful at informing therapeutic decisions.However, genomics remains relatively limited in its ability to predict the onset of most complex diseases, largely because genomic information does not account for dynamic environmental influences (1).To better understand lung disease, one needs to examine the downstream changes occurring at the level of proteins and metabolites.
Proteins are the main effectors of cellular physiology.Therefore, the proteome, alone or through its integration with other systems sciences, is a particularly informative tool for understanding pulmonary diseases (Figure 1) (2).Until recently, technology was limited to studying the role of single proteins.Although mass spectrometry (MS) has been an important tool for decades, new technologies and strategies in peptide/ protein separation, MS analysis, quantitative protein analysis, and databases now enable the simultaneous analysis of dozens to even thousands of proteins in a single biological sample.In parallel, advances in statistical and bioinformatics tools now allow insight into protein pathways and networks involved in lung disease.Consequently, there has been a surge in the number of proteomics publications related to lung disease (Figure 2) (3).
Metabolites, small biological compounds with a low molecular weight (typically <1,500 Daltons), reflect the activity of proteins, and serve as signaling molecules for processes that include gene and protein regulation (4, 5) (Figure 1).As such, the added value of metabolomics (i.e., the simultaneous measurement of small molecules in a biological sample) is that it reflects, complements, and informs data acquired by other systems biology sciences (Figure 1) (6).Because metabolic profiles change rapidly with the biologic state, metabolomics permits unique insight into both the pathogenesis of disease and drug response (pharmacometabolomics), as well as often unapparent phenotypes and endotypes.The trans-omic approach provides a unique opportunity to gain insights in to how genetic

AMERICAN THORACIC SOCIETY DOCUMENTS
programs are translated into biological function and how alterations in the program associate with the onset of diseases.As such, metabolomics may reveal early biomarkers that could improve risk assessment and diagnosis of complex diseases.Furthermore, metabolic profiles are influenced by exogenous factors, including medications, lifestyle, and the environment, so metabolomic profiling has the potential to unravel the impact of both genetic and nongenetic factors on disease onset, progression, and severity.This report highlights the significant progress and continued major challenges in the fields of proteomics and metabolomics, with particular focus on lung health and disease.

Methods
The conference was created by the American Thoracic Society Respiratory Cell Molecular Biology Assembly Working Group on Proteomics and Metabolomics.The goal of the workshop was to convene for the first time relevant experts in lung proteomics and metabolomics to discuss specific challenges that are unique to the lung.The main objectives were to: d Identify existing and emerging technologies in metabolomics and proteomics as applied to the study of the lung.d Understand the inherent challenges associated with metabolomics and proteomics, with a specific focus on challenges associated with lung biological specimens.d Identify informatics approaches and online databases relating to metabolomics and proteomics.d Discuss systems biology approaches, including integrative platforms across databases that can be applied to metabolomic and proteomic studies.d Discuss the potential clinical applications of proteomics and metabolomics for lung-related disorders.
The conference included combined sessions relevant to proteomics and metabolomics, and breakout sessions highlighting challenges specific to each.The conference report is a summation of the presentations and discussions.Potential conflicts of interest were disclosed and managed in accordance with the policies and procedures of the American Thoracic Society.

Issues Common to Proteomics and Metabolomics Unique Challenges of Lung Biospecimens
The lung has several features that make metabolomics and proteomics both a unique opportunity and challenge compared with other organs and blood.One unique quality is the lung's exposure to the environment (air pollution, pollen, etc.); however, as part of an environmental defense, the lung is covered by a complex epithelial defensive barrier and epithelial lining fluid (ELF) consisting of both solute and gel phases containing mucus and lipids (e.g., surfactants), along with inflammatory cells.These features make sample preparation challenging.The collection of lung-specific specimens includes tissue biopsy or pathology specimens, transbronchial biopsies, bronchoalveolar lavage fluid (BALF), and airway brushings.The most frequent lung sampling technique is bronchoscopy to obtain BALF and/or bronchial biopsy.Of note, as in all tissues, biopsies lead to blood content contamination that may influence sample analysis.This may be less of an issue for airway wall biopsies.BALF is typically performed using normal saline (0.9% NaCl), often requiring techniques to remove the high salt content.Furthermore, BALF dilutes the ELF up to 100-fold, and must be taken into account when performing quantitative analysis.Dilution can be minimized by using a single-cycle lavage and corrected by normalizing to urea (7, 8).ELF is rich in plasma-derived proteins (albumin, transferrin, etc.), along with proteins specifically expressed by airway cells, such as surfactant proteins and club cell secretory protein (9).
Other less frequently used biosamples for lung investigations include exhaled breath condensate and epithelial brushes or biopsies.The former is limited by difficulty in standardization and dilute samples, and the latter by bleeding and invasiveness of procedures.Sputum is another lung sample with its own unique challenges, such as viscosity and different layers (gel and sol).To minimize variance and maximize reproducibility, we have included in this document recommended sample protocols for lung and other biofluid samples (Table 1).AMERICAN THORACIC SOCIETY DOCUMENTS For proteomics, it is difficult to standardize; very dilute specimens (,1 µg/ml protein).
Lung biopsy or transbronchial biopsy After obtaining samples they are snap frozen in liquid nitrogen; 1-5 g of tissue is processed using a bead-based tissue maceration method (such as a genogrinder).Once this is done, one can use the usual preparation metabolomics methods for NMR, GC-MS, or LC-MS as used for liquid samples.

Bronchial wall brushings
Again, after obtaining samples they are snap frozen in liquid nitrogen.It is best if 1-5 g or the equivalent of 10,000 cells are processed using a bead-based tissue maceration method (such as a genogrinder).Once this is done, one can use the usual preparation metabolomics methods for NMR, GC-MS or LC-MS as used for liquid samples.
The sample can be placed in saline and then centrifuged, snap frozen and then processed.

Plasma
Collect blood by direct venipuncture, if possible, into a vacutainer tube containing either EDTA or sodium heparin.Immediately invert the tube several times to ensure mixture with anticoagulant.Within 30 min of blood collection, centrifuge balanced tubes (15 min at 1,300 3 g) with no brake to ensure proper plasma separation.Refrigeration before or during centrifugation is recommended for metabolomic studies but not recommended for proteomic studies.After centrifugation, the blood should be separated into 3 visible layers, the upper layer is generally clear and pale yellow in color and is the plasma.The second, thin, whitish layer sits at the interface between the plasma and the red blood cells, and is called the buffy coat.For proteomics, consider the addition of protease inhibitors in blood draw tube (e.g., BD P100); in general, this is not recommended for all tubes and is not recommended for metabolomics. (Continued)

AMERICAN THORACIC SOCIETY DOCUMENTS
Invasive lung samples are not often readily accessible, so blood (whole blood, serum, plasma), urine, and isolated cells (e.g., airway epithelial brushings or alveolar macrophages) are often used to indirectly study the lung proteome and metabolome.Although these are more readily accessible, their relevance to lung disease is often less clear.

Shared Resources
Dr. Shankar Subramaniam (University of California, San Diego, CA) discussed the challenges of shared databases, repositories, and software for metabolomics studies.The National Institutes of Health (NIH, Bethesda, MD) supports resources to organize and store national and international metabolomics data and analysis tools through the Metabolomics Workbench program (http://www.metabolomicsworkbench.org//nihmetabolomics/index.html),housed and managed by the University of California San Diego Supercomputer Center.In addition, the site is a resource for analytical standards important for confirming metabolite identities and lipid map classifications.On the horizon is the expected publication of the NIH-sponsored Ring Trial in Metabolomics, in which several metabolomics laboratories across the United States assayed technical replicates of samples to assess standardized processes, analyte detection, and data reproducibility across centers.

Data Integrity
Dr. Arthur Moseley (Duke University, Durham, NC) discussed the importance of project-to-project and laboratory-tolaboratory reproducibility in proteomics and metabolomics.Although not unique to the lung, the lack of standardization in proteomics and metabolomics analyses presents unique challenges to data integrity.There are strategies to improve standardization.As an example, the platebased targeted metabolomic platform (Biocrates Absolute IDQ p180; Innsbruck, Tirol, Austria) uses internal standards and calibration curves for precise metabolite quantitation, and has been validated across all major MS vendors.Similarly, quality control pools containing a mixture of studyspecific samples and reference standards (e.g., human plasma from Golden West Biologicals, Inc., Temecula, CA) can be used to measure intra-and interstudy reproducibility.These approaches can help to overcome batch effects, and should ensure that identical results can be achieved across laboratories and instrument platforms.However, ultimately, the harmonization of analytical and quality control methods will improve and ensure metabolomic and proteomic data integrity (10, 11).

Statistical Approaches
Careful analysis of complex data is essential to fully capture all potential opportunities to explore biological systems and disease.Although proteomics and metabolomics enable accurate detection and quantification in an unbiased manner, there are some unique statistical challenges in assessing these data.Dr. Katerina Kechris (University of Colorado, Denver, CO) discussed how untargeted MS, by definition, does not include chemical standards and, therefore, measurements reflect relative abundances.When studies are completed over multiple experiments, there is frequently drift in retention time (RT) and sensitivity requiring batch correction by using methods, such as Combat or Remove Unwanted Variation (12-14).Metabolites, in particular, tend to be highly correlated within a class.This requires class analysis or the use of dimension reduction through principal component analysis, partial least squares projection to latent structures, clustering, or other multivariate methods (15).Correction for multiple comparisons using methods, such as the false discovery rate, is essential, because many proteins and metabolites can be simultaneously tested (16).In addition, proteins and metabolites are most often not normally distributed and require data transformations (e.g., logarithmic) to normalize the data.It is also important to adjust for covariates, such as age, sex, and smoking history, that can influence metabolite and protein expression.Finally, unique challenges include handling missing values, mass spectra acquisition, identification, and multiple sources of variability (17).

Current State of Metabolomics of Lung Diseases
The fundamental premise in metabolomics is that changes, whether physiological or pathological, cause alterations of the metabolome that are detected as variations

Challenges in Metabolomics
Major challenges in metabolomics include data processing, identification and validation of metabolites, and data visualization.
Importantly, because no one analytical

AMERICAN THORACIC SOCIETY DOCUMENTS
American Thoracic Society Documents platform captures the entire metabolome, there are a number of analytical platforms that can be used (Figure 3), each of which has its own advantages and disadvantages (Table 3).An overall consensus was a call for greater synchrony of research methods to enable development of "big data" resources shared across institutions that can be applied to lung diseases.

Metabolite Identification
MS data are often reported as features, which represent analytes with discrete massto-charge ratios (m/z) and RTs.Dr. Richard Reisdorph (University of Colorado, Denver, CO) discussed the importance of metabolite identification, rather than a simple description of these features, as many represent breakdown products of known metabolites.Differential features are identified, and extensive follow-up work is needed to identify the specific compound.This is particularly challenging for lipids, as large numbers of lipids have very similar or identical m/z and RT.Gas chromatography annotation libraries are more developed than liquid chromatography (LC) libraries, but, to date, a standard metabolite library for untargeted LC-tandem MS (MS/MS) is not readily available for all researchers.This limits downstream pathway analysis, which can only be as good as the annotation of known metabolites.To move the field forward, a description of the annotation confidence is important, ranging from low confidence (e.g., mass of analyte matched to a database) to high (MS/MS spectrum matches to an MS/MS library) to highest "gold standard" (confirmed with purchased standards, RT, and MS/MS spectrum).To this end, there is a need for analytical reference standards for metabolomics (http://www.metabolomicsworkbench.org/ standards/index.php).Data validation, as reviewed by Dr. Nichole Reisdorph (University of Colorado, Denver, CO), with a targeted assay in an independent population is the gold standard, but requires a substantial investment of time to confirm metabolite identifications.

Data Processing
Data processing needs, which differ significantly for nuclear magnetic resonance (NMR) spectroscopy and MS approaches, were discussed by Dr. Dean Jones (Emory University, Atlanta, GA).The analysis of NMR spectra can be challenging due to peak overlap, but analysis can be optimized with the use of software that permits the identification and quantification of metabolites (6).Spectral peak overlap is more readily deconvoluted for LC-MS than for NMR.For both sources of data, chemometric methods can be used in which peak or feature signals

Cutting-Edge Metabolomics
Advanced Imaging Using MS Technology Dr. Richard Caprioli (Vanderbilt University, Nashville, TN) described how matrixassisted laser desorption/ionization imaging MS produces molecular maps of peptides, proteins, lipids, and metabolites present in intact tissue sections (29,30).This technique employs desorption of molecules by direct laser irradiation to map the location of specific molecules from freshfrozen and formalin-fixed tissue sections without the need of target-specific reagents, such as antibodies.Molecular images of this nature are produced in specific m/z values, or ranges of values.Each specimen gives rise to many hundreds of specific molecular images from a single raster of the tissue.In a complementary approach, where only discrete areas within the tissue are of interest, a histology-directed approach that integrates MS and microscopy has been developed.Thus, mass spectra are collected from only selected areas of cells within the tissue after laser ablation and analysis.
Clinically relevant studies include advanced diabetic nephropathy involving both proteins and lipids (29).In addition, imaging MS has been applied to drug targeting and metabolic studies, both in specific organs and also in intact whole animal sections after drug administration.These techniques, though promising, have yet to be applied to the lung outside of lung cancer.

The Future of Lung Metabolomics
Dr. Brent Winston (University of Calgary, Calgary, AB, Canada) discussed the key strategies for advancing lung metabolomics.Data sharing and adopting standards of practice (SOPs) are key for future studies.In addition, a focus on targeted mechanistic studies will enable the field to move beyond hypothesis-agnostic discovery science.The

AMERICAN THORACIC SOCIETY DOCUMENTS
American Thoracic Society Documents AMERICAN THORACIC SOCIETY DOCUMENTS The use of existing samples from major lung-related studies is highly desirable, but is only possible if samples are collected with proper SOPs and appropriate storage of samples in an internationally recognized fashion (32).This will require cooperation from major funding agencies for all phase 1, 2, and 3 studies involving the respiratory system.For example, the Precision Medicine Initiative (https://ghr.nlm.nih.gov/primer/precisionmedicine/initiative) and the Million Veteran Program (http:// www.research.va.gov/mvp/) plan to collect samples from one million subjects, each using uniform methodology.To do this well would include the development and harmonization of SOPs, storage methods, sample annotation, data sharing, and exploration of age-related storage material degradation (33,34).

Current State of Proteomics in Lung Diseases
Proteomic Approaches in Studying Lung Disease MS has revolutionized the study of proteins, as it allows the measurement of hundreds to thousands of proteins in complex systems in a very precise and reproducible manner.In the first presentation of the proteomics breakout session, Dr. Chris Wendt (University of Minnesota, Minneapolis, MN) discussed the current state of proteomics in lung disease, specifically focusing on ARDS, COPD, and idiopathic pulmonary fibrosis (Table 7).Many of these studies used two-dimensional gel electrophoresis and/or a combination of LC and MS, common techniques for the identification of disease biomarkers and disease-related signaling pathways using gene ontology analysis (35)(36)(37)(38)(39). Specific advances and resources in databases, repositories, methods, and protocols have greatly advanced the field of proteomics (Table 8); however, challenges remain.The MAIT package contains functions to perform endto-end statistical analysis of LC-MS metabolomics data.Special emphasis is put on peak annotation and in modular function design of the functions. (Continued)

AMERICAN THORACIC SOCIETY DOCUMENTS Challenges in Quantitative Proteomics
Challenges remain in the ability to accurately quantify changes in protein abundance.Both label and label-free strategies exist for proteome quantitation (Table 9).Dr. Alexey Nesvizhskii (University of Michigan, Ann Arbor, MI) discussed the application of label-free quantitative methods, including software tools for data-dependent acquisition with quantitation by spectral counting or ion abundance (e.g., QSpec/QProt) (40,41), as well as data-independent acquisition (DIA; e.g., DIA-Umpire) (42, 43) (Figure 4).Spectral counting is defined as the number of MS/MS sequencing attempts made on a precursor (i.e., intact peptide) during a single LC-MS/MS analysis, whereas intensity-based quantitation is the measurement of the area under the curve of each precursor in a sample.Intensity-based methods have greater accuracy, but both methods suffer from missing data across replicate analyses.The missing data problem can be mostly eliminated by aligning (or matching) of precursors across batched analyses (44).
Data-Independent Methods for Quantitative Proteomics DIA methods use "MS2-based" quantitation of the ions that are produced by MS/MS fragmentation (45) and can have greater selectivity (and greater signal-to-noise) than precursor/MS1based quantitation.DIA also allows for matching of each sample to an external library of MS/MS spectra, thus largely eliminating the missing data problem.DIA has shown promise for analysis of biofluids, including BALF (46).DIA-Umpire software can furthermore extract MS/MS spectra for peptide/protein identification using conventional database searching (42,43).
Quantifying the Secretome Proteome LC-MS/MS allows the identification and quantification of hundreds of proteins in cellular secretions (secretomes) of airway cells.This includes the analysis of airway cells in vitro, which allows for the identification of key mechanistic biochemical insights, and thus plays a pivotal role in translational lung research.However, there remain challenges in quantifying the secretome, as discussed by Dr. Kristy Brown (Children's National Health System, Washington, D.C.) and Dr. Mehmet Kesimer (University of North Carolina-Chapel Hill, Chapel Hill, NC).Dr. Brown studies the altered secretome of human bronchial epithelial cells (HBECs) obtained from patients with cystic fibrosis (47,48).Here, she introduced the concept of stable isotope labeling of amino acids in cell culture (SILAC), which uses cells grown in isotopically labeled amino acids (typical 13 C-and 15 N-labeled arginine, lysine, and/or leucine ) to synthesize the "heavy" forms of proteins that can be mixed with their "light" counterparts before trypsinization and/or peptide/protein fractionation (Figure 5).Advantages of SILAC include accuracy in quantitation and the ability to simultaneously quantify and differentiate two proteomes.This method is also well suited to studying proteome-wide protein synthesis and decay, and a "SuperSILAC" mix, such as that generated from ARDS secretions, can be used as a common reference standard (49).This approach could have general utility for airway secretomics; however, SILAC techniques usually require numerous passages with media containing the heavy amino acids, limiting its applicability to cells in culture.
Label-Free Quantitative Proteomics Dr. Kesimer emphasized the utility of label-free quantitation of lung secretions.In a typical label-free proteomic experiment, samples are normalized to total protein before trypsinization, and are intensity normalized during data analysis.He discussed the potential pitfalls of this approach for comparing health to disease (e.g., how large increases in mucin-5B secretion by HBECs were compressed when data were intensity normalized across all samples).Urea has been long used as a normalizing factor in protein quantitation of BALF.It stands to reason that a similar approach (normalization to an "unperturbed" metabolite or protein) might improve quantitative accuracy of HBEC secretomics.

Proteomic Approaches to Cell Signaling
As is the case with epithelial cells, proteomic evaluation of isolated myeloid cells has the virtue of providing focused molecular insights into physiologic events that occur in vivo in the multicellular lung.In particular, ex vivo culture of sentinel immune cells, such as macrophages, allows for study of time-resolved cell signaling events induced by environmental stimuli, such as bacterial LPS.Dr. Michael Fessler (National Institute of Environmental Health Sciences, NIH) discussed proteomic approaches to cell signaling using primary murine macrophages (50), primary human neutrophils (51), and immortalized macrophage cell lines (52,53).Whereas phosphoproteomic strategies partnered with chemical inhibitors or RNA interference can be used to map out kinase cascades, subcellular fractionation and immunoprecipitation can permit focused insight into compartmentalization of signaling events within the cell.SILAC has been applied successfully to both primary (e.g., bone marrow-derived) and immortalized macrophages, and can be used in multiple signaling applications, including detecting changes in posttranslational modification, localization, and interaction of signaling proteins.SILAC has been particularly valuable in kinetic studies (e.g., measurement of protein turnover) and analysis of signaling events (e.g., phosphorylation).In addition, labelfree approaches, such as spectral counting, can be used in a semiquantitative manner to monitor targeted signaling events within the cell, and can also be used to help validate specificity in pulldown assays.

AMERICAN THORACIC SOCIETY DOCUMENTS
American Thoracic Society Documents The application of LC-MS/MS for the identification and quantitation of select peptides/proteins (targeted proteomics) is an alternative to immunoassay-based protein quantitation (10, 54).Although immunoassays may have greater sensitivity, targeted proteomic assays use internal standards to achieve high specificity, offer a high degree of multiplexing, and enable facile quantitation of post-translational modifications.These advantages were discussed in a presentation by Dr. Matt Foster (Duke University Durham, NC), which also served as an introduction to the design and application of targeted proteomic assays to airway cells and biofluids.Dr. Foster has employed targeted proteomic assays for the quantitation of cytokines and chemokines in BALF (55), and the quantitation of allelic variants and isoforms of surfactant protein A (56).Additional applications include quantitation of genetic lineages of human metapneumovirus from cell culture and nasal lavage specimens (57).Targeted proteomics is also a powerful tool for quantitation of post-translationally modified peptides beyond phosphorylation (e.g., small ubiquitin-like modifier (SUMO) modification, methylation, acetylation, ubiquitinylation, acylation, and oxidation) that can now be measured by the thousands in discovery-based proteomics studies.To this end, he presented data on a targeted proteomic assay for quantitation of newly discovered phosphorylation sites in basal cell cytokeratins (58).Targeted proteomic assay development has been a major focus of the NIH-funded Clinical Proteomic Tumor Analysis Consortium (CPTAC) (10), and this technique will likely have an important future role in clinical diagnostic and prognostic assays for lung diseases.

Summary Integration
This symposium focused on current challenges in applying emerging metabolomics and proteomics methodologies to lung disease.Unlike genetics and genomics, where advanced sequencing technology allows independent laboratories to achieve highly similar results, metabolomics and proteomic profiling remains challenging for several key reasons.First, there are the inherent challenges to metabolomics and proteomics, such as identification and quantification of both peptides and metabolites, along with "big data" analyses that aggregate samples and data across many laboratories and impair the feasibility of systems biology data integration from multiple sources.In addition, there are lung-specific issues, such as lack of uniform SOPs specifically for the lung, leading to operator and protocol variability with sample attainment.Working to advocate for an NIH/NHLBI investment in publicly available, well phenotyped biobanks that include diverse sample types (e.g., plasma, BALF, lung biopsies) would help standardize proteomic and metabolomic methods and further the field of biomarker development in lung diseases.This emerged as a major goal of this symposium.
Although there may be challenges to standardize large population metabolomics and proteomics analyses across different platforms, the potential benefit for

AMERICAN THORACIC SOCIETY DOCUMENTS
American Thoracic Society Documents overcoming these challenges can be illustrated through more focused approaches on single platforms by individual laboratories.For example, Dr. Michael Snyder (Stanford University, Stanford, CA) highlighted his own experience using a longitudinal integrative personal omics profile (iPOP), examining genomic, transcriptomic, proteomic, metabolomic, and autoantibody profiles from a single individual over a 14-month period (59).This strategy revealed how iPOP can detect prediabetes as well as routine viral infections.This type of approach in which proteomics and metabolomics are integrated with other "omics" at the individual level is technically feasible and is an outstanding example of precision medicine; however, there remain cost and bioinformatic challenges that need to be overcome before the iPOP becomes routine in clinical practice.
In conclusion, the symposium demonstrated how proteomics and metabolomics can be used to better understand and track lung diseases, but there  In a typical unbiased proteomic analysis, tryptic peptides are separated using liquid chromatography and introduced into the mass spectrometer using electrospray ionization.The sum intensity of detected peptides is often visualized as total ion current over time, as in A. (B) In a typical cycle of a data-dependent analysis (DDA), a "full scan" of all precursor (MS1) ions present is performed followed by (C) tandem mass spectrometry (MS/MS) analysis of the topN (e.g., top3; starred peaks) most abundant ions.The MS/MS spectra are used for database searching to identify the corresponding peptides.(D) Finally, identified peptides are quantified based on the area under the curve (AUC) of the MS1 intensity.(E) In a dataindependent analysis (DIA), all ions within a selected mass range are subjected to MS/MS fragmentation.(F) Quantitation is performed by AUC of the fragment ions (MS2) that belong to a particular peptide.m/z = mass-to-charge ratio; MS1 = mass spectrometry analyzer 1; MS2 = mass spectrometry analyzer 2; TIC = total ion chromatogram; XIC = extracted ion chromatogram.

Figure 2 .
Figure 2. Temporal increase in the number of lung proteomics and metabolomics publications in the PubMed database.Squares represent proteomic publications; triangles represent metabolomic publications.

Figure 3 .
Figure 3.The most commonly used analytical platforms for metabolomics are: (A) proton ( 1 H) nuclear magnetic resonance (NMR); (B) gas chromatography (GC)-mass spectroscopy (MS); and (C) liquid chromatography (LC)-MS.(A) NMR is ideal for the detection of polar compounds like amino acids and for smaller molecular weight (<100 Da) metabolites that LC-MS can miss.NMR is routinely quantitative when an internal standard, such as 4,4-dimethyl-4-silapentane-1sulfonic acid (DSS), is added to the sample.(B) GC-MS most often requires the derivatization of volatile compounds that are separated by a gas carrier phase and elute based on retention time in the column.After ionization, compounds are detected by MS.The graphic printout shows a typical serum readout of abundance versus time (top) and abundance versus mass/charge ratio (bottom).(C) For LC-MS metabolomics, molecules are ionized, typically by electrospray ionization, and the resulting positive and negative ions are detected by MS.This results in a mass-to-charge ratio (m/z) versus relative peak intensity graphical representation of the data.More details about the advantages and disadvantage of each approach can be found in Table 2.By K. Murray (Kkmurray) (Own work) [GFDL (http://www.gnu.org/copyleft/fdl.html),CC-BY-SA-3.0 (http://creativecommons.org/licenses/by-sa/3.0/)or CC BY-SA 2.5-2.0-1.0 (http://creativecommons.org/licenses/by-sa/2.5-2.0-1.0)],via Wikimedia Commons.

109
Definition of abbreviations: 1D-3D = one to three dimensional; CluPA = cluster-based peak alignment; GC = gas chromatography; LC = liquid chromatography; MCMC = Markov chain Monte Carlo; MS = mass spectroscopy; NMR = nuclear magnetic resonance.*Information about instrument specific software packages can be found at each respective manufacturer's website.
remain important necessary steps to bring these fields into routine clinical practice.n This official Workshop Report was prepared by an ad hoc subcommittee of the American Thoracic Society Workgroup on Metabolomics and Proteomics.Members of the Subcommittee are as follows: RUSSELL P. BOWLER, M.D., PH.D. (Co-Chair) CHRIS H. WENDT, M.D. (Co-Chair) MICHAEL B. FESSLER, M.D. MATTHEW W. FOSTER, PH.D. RACHEL S. KELLY, M.P.H., Ph.D. JESSICA LASKY-SU, Sc.D.

Figure 4 .
Figure 4. Summary of two common proteomic data acquisition methods.In a typical unbiased proteomic analysis, tryptic peptides are separated using liquid chromatography and introduced into the mass spectrometer using electrospray ionization.The sum intensity of detected peptides is often visualized as total ion current over time, as in A. (B) In a typical cycle of a data-dependent analysis (DDA), a "full scan" of all precursor (MS1) ions present is performed followed by (C) tandem mass spectrometry (MS/MS) analysis of the topN (e.g., top3; starred peaks) most abundant ions.The MS/MS spectra are used for database searching to identify the corresponding peptides.(D) Finally, identified peptides are quantified based on the area under the curve (AUC) of the MS1 intensity.(E) In a dataindependent analysis (DIA), all ions within a selected mass range are subjected to MS/MS fragmentation.(F) Quantitation is performed by AUC of the fragment ions (MS2) that belong to a particular peptide.m/z = mass-to-charge ratio; MS1 = mass spectrometry analyzer 1; MS2 = mass spectrometry analyzer 2; TIC = total ion chromatogram; XIC = extracted ion chromatogram.

Table 1 .
Sample preparation for proteomics and metabolomics studies*

Table 3 .
Exa32)es of strengths and weaknesses of nuclear magnetic resonance and MS methods used in metabolomics studies(6,32) Definition of abbreviations: GC = gas chromatography; LC = liquid chromatography; MS = mass spectroscopy; NMR = nuclear magnetic resonance.Adapted from Reference 32.

Table 4 .
Publicly available tools for metabolite identification* The database contains .42,000metabolite entries, including chemical/clinical/enzymatic data, and links to proteins and other databases (KEGG, PubChem, MetaCyc, ChEBI, PDB, UniProt, and GenBank) and a variety of structure and pathway viewing applets.The HMDB database supports extensive text, sequence, chemical structure, and relational query searches.Four additional databases, DrugBank, T3DB, SMPDB, and FooDB, are also part of the HMDB suite of databases.91 *Information about instrument specific software packages can be found at each respective manufacturer's website.

Table 5 .
Publicly available and commercial software/tools for spectral processing and analysis for metabolomics*

Table 6 .
Publicly available and commercial software/tools for analysis and interpretation of metabolomics data*

Table 6 .
(Continued )Definition of abbreviations: LC = liquid chromatography; MS = mass spectroscopy; MSEA = metabolite set enrichment analysis *Tools such as XCMS and MetSign in TableIinclude data analysis and visualization options.

Table 7 .
Proteomic biomarker publications in acute respiratory distress syndrome, chronic obstructive pulmonary disease, and idiopathic pulmonary fibrosis

Table 9 .
Overview of quantitation strategies for shotgun proteomic analyses PRM Ratio of AUC for native versus SIL internal standard.Can be applied to large panels or to a few select targets.