Combination Strategy of Reactive and Catalytic Matrices for Qualitative and Quantitative Profiling of N-Glycans in MALDI-MS

In Situ Glycan Analysis and Editing in Living Systems

Besides proteins and nucleic acids, carbohydrates are also ubiquitous building blocks of living systems. Approximately 70% of mammalian proteins are glycosylated. Glycans not only provide structural support for living systems but also act as crucial regulators of cellular functions. As a result, they are considered essential pieces of the life science puzzle. However, research on glycans has lagged far behind that on proteins and nucleic acids. The main reason is that glycans are not direct products of gene coding, and their synthesis is nontemplated. In addition, the diversity of monosaccharide species and their linkage patterns contribute to the complexity of the glycan structures, which is the molecular basis for their diverse functions. Research in glycobiology is extremely challenging, especially for the in situ elucidation of glycan structures and functions. There is an urgent need to develop highly specific glycan labeling tools and imaging methods and devise glycan editing strategies. This Perspective focuses on the challenges of in situ analysis of glycans in living systems at three spatial levels (i.e., cell, tissue, and in vivo) and highlights recent advances and directions in glycan labeling, imaging, and editing tools. We believe that examining the current development landscape and the existing bottlenecks can drive the evolution of in situ glycan analysis and intervention strategies and provide glycan-based insights for clinical diagnosis and therapeutics.

α-Cyano-3-aminocinnamic acid: A novel reactive matrix for qualitative and quantitative analysis of plant N-glycans by MALDI-MS

N-glycans have a diversity of crucial biological roles in organisms. Matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) has become an indispensable analytical instrument for biomolecules. However, due to the inherent low abundance, high structural heterogeneity, and poor ionization efficiency of N-glycans, as well as the extremely inhomogeneous co-crystal property using traditional matrices, the qualitation and quantitation of N-glycans by MALDI-MS remains challenging. In the present study, α-cyano-3-aminocinnamic acid (3-CACA) was reasonably designed and synthesized as a novel reactive matrix for N-glycan analysis. Combining with traditional matrix α-cyano-4-hydroxycinnamic acid (CHCA) as an acidic catalyst, a combinational matrix 3-CACA/CHCA was obtained with homogeneous co-crystallization and high derivatization efficiency, achieving the sensitive qualitation with the limits of detection low to femtomole and reproducible quantitation with good linearity (R2 > 0.998). As a result, the established method was successfully applied to the on-target derivatization and high-throughput quantification of N-glycans in eight varieties of the peach complex system, indicating that N-glycan has the potential to become a new biomarker for food allergy, and elucidating the prospective correlation between N-glycan epitopes and allergic reactions.

Collision Induced Unfolding Enables the Quantitation of Isomass Biotherapeutics in Complex Biological Matrices

Quantitative mass spectrometry has been widely used to evaluate the concentrations of molecules within a variety of biological matrices. Typically, such quantitative mass spectrometry analyses are predicated upon the production of mass-resolved precursor or fragment ions, leading to challenges surrounding the quantification of isomeric or conformationally distinct analytes. As such, new approaches are required for the label-free quantification of isomass proteins. Native ion-mobility MS (nIM-MS) in combination with collision induced unfolding (CIU) is a potentially enabling approach for such quantitative mass spectrometry methods as the technique can rapidly separate and detect many biomacromolecule isoforms. CIU uses collisional activation to capture the unfolding trajectory of ions in the gas phase, producing different intermediate structures that can be leveraged to distinguish protein structures that exhibit identical sizes at lower energies. Here we describe the deployment of quantitative CIU methodology to measure the concentrations of isomass pairs of biotherapeutics and sequence homologues in both standard and biological matrices. Our results cover three antibody pairs and include examples of mixed therapies where multiple biologics are commonly provided to patients. In all cases, CIU enables the production of resolved features for each antibody mixture probed, producing calibration curves with correlation coefficients ranging from 0.92 to 0.99, limits of detection ranging from 300 to 5000 nM and sensitivities ranging from 8.7 × 10-5 nM-1 to 6 × 10-3 μM-1. We conclude our report by projecting the future utility of CIU-enabled quantitative MS methods.

Quantitative N-Glycomic and N-Glycoproteomic Profiling of Peach [Prunus persica (L.) Batsch] during Fruit Ripening

Being part of the human diet, peach is an important fruit consumed worldwide. In the present study, a systematic first insight into the N-glycosylation of peach fruit during ripening was provided. First, N-glycome by reactive matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectrometry indicated that 6 of 24 N-glycans of peach were differentially expressed. Second, a comparative N-glycoproteome was characterized via ¹⁸O-tagged N-glycosylation site labeling followed by nano-liquid chromatography-electrospray ionization-tandem mass spectrometry (nLC-ESI-MS/MS). Totally 1464 N-glycosites on 881 N-glycoproteins were identified, among which 291 N-glycosites on 237 N-glycoproteins were expressed differentially with a fold change value of 1.5 or 0.67. The enrichment analysis of GO and KEGG revealed that four pathways including other glycan degradation, phenylpropanoid biosynthesis, amino sugar and nucleotide sugar metabolism, and protein processing in endoplasmic reticulum were mainly enriched, in which several important N-glycoproteins with dynamic change during fruit ripening were further screened out. Our findings on a large scale for N-glycosylation analysis of peach fruit during ripening may provide new molecular insights for comprehending N-glycoprotein functions, which should be of great interest to both glycobiologists and analytical chemists.

2-Hydrazinoterephthalic Acid as a Novel Negative-Ion Matrix-Assisted Laser Desorption/Ionization Matrix for Qualitative and Quantitative Matrix-Assisted Laser Desorption/Ionization-Mass Spectrometry Analysis of N-Glycans in Peach Allergy Research

Glycans recently attracted considerable attention as the proposal of cross-reactive carbohydrate determinants for food allergy. Matrix-assisted laser desorption/ionization-mass spectrometry (MALDI-MS) is powerful in analyzing biomolecules, while its applications in glycans are still challenging. Herein, a novel reactive matrix-assisted laser desorption/ionization (MALDI) matrix, 2-hydrazinoterephthalic acid, was rationally designed and synthesized. It provides uniform co-crystallization with glycans and only produces deprotonated ions with high intensities in the negative-ion mode. In combination with sinapic acid, a rapid and high-throughput method was established for on-target analysis of glycans with a superior limit of detection at the femtomole level and a good linearity (R² > 0.999). Furthermore, the established method was successfully applied to quantify N-glycans in different cultivars and tissues of peach [Prunus persica (L.) Batsch]. Our work suggests the potential role of N-glycans as biomarkers for food-borne allergy and lays a methodological foundation for the elucidation of the possible relationship between carbohydrate epitopes and food allergy.

Hydrophilic, dual amino acid-functionalized zinc sulfide quantum dot for specific identification of N-glycopeptides from biological samples

RATIONALE

Glycosylation of proteins is one of the most significant and complex post-translational modifications, and N-glycosylation plays a crucial role in life activities. Mass spectrometry (MS) has been a powerful technique in the analysis of protein glycosylation. However, the direct detection of glycoproteins in biological samples based on MS still suffers from huge challenges. Therefore, enrichment and purification of samples before MS analysis is an essential prerequisite.

METHODS

Hydrophilic interaction liquid chromatography (HILIC) has significantly developed for selective enrichment of glycopeptides due to its simple operation process and unbiased enrichment. Herein, hydrophilic, dual amino acid-functionalized zinc sulfide quantum dots (ZnS QDs) were prepared to enrich glycopeptides using an easy procedure. The enriched glycopeptides were detected using matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS).

RESULTS

The obtained material exhibited high selectivity (1:2000), low detection limit (0.1 fmol/μl), good repeatability (10 times), and excellent recovery (89.8%) in glycopeptide enrichment. In the actual application in biological samples, 71 N-glycopeptides and 161 N-glycopeptides were detected from human saliva and serum, respectively.

CONCLUSIONS

ZnS-Au-GC was successfully prepared using an easy method. The results showed that the obtained material exhibited excellent performance in glycopeptide enrichment. Furthermore, it had showed great potential for glycopeptide enrichment in complex biological samples.

Analysis of carbohydrates and glycoconjugates by matrix-assisted laser desorption/ionization mass spectrometry: An update for 2019-2020

This review is the tenth update of the original article published in 1999 on the application of matrix-assisted laser desorption/ionization (MALDI) mass spectrometry to the analysis of carbohydrates and glycoconjugates and brings coverage of the literature to the end of 2020. Also included are papers that describe methods appropriate to analysis by MALDI, such as sample preparation techniques, even though the ionization method is not MALDI. The review is basically divided into three sections: (1) general aspects such as theory of the MALDI process, matrices, derivatization, MALDI imaging, fragmentation, quantification and the use of arrays. (2) Applications to various structural types such as oligo- and polysaccharides, glycoproteins, glycolipids, glycosides and biopharmaceuticals, and (3) other areas such as medicine, industrial processes and glycan synthesis where MALDI is extensively used. Much of the material relating to applications is presented in tabular form. The reported work shows increasing use of incorporation of new techniques such as ion mobility and the enormous impact that MALDI imaging is having. MALDI, although invented nearly 40 years ago is still an ideal technique for carbohydrate analysis and advancements in the technique and range of applications show little sign of diminishing.

On-tissue derivatization for isomer-specific mass spectrometry imaging and relative quantification of monosaccharides in biological tissues

Spatially-resolved profiling of tissue monosaccharides not only gives an insight into the spatial heterogeneity of monosaccharides, but also helps to decipher the possible roles of monosaccharides in biological processes. Here, we develop an on-tissue derivatization method, coupled with matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) to image and quantify the aldose and ketose isomers of monosaccharide in biological tissues. A new derivatization reagent, 1-naphthaleneacethydrazide (NAH) was synthesized for the on-tissue derivatization of monosaccharides, and it significantly enhanced the imaging sensitivity of monosaccharides. Moreover, the NAH-derivatized aldose and ketose can generate isomer-specific diagnostic ions during MALDI-MS/MS analysis, and thus paves way for the isomer-specific MS imaging of aldose and ketose monosaccharides. On this basis, we further constructed a quantitative MALDI-MS imaging model based on isomer-specific diagnostic ions, and calculated the expression contents of aldose and ketose monosaccharide isomers in different tissue regions of carrot section. We expect that the development of this method should provide more precise view on the spatial distributions and contents of different monosaccharides in heterogeneous biological tissues.

Metabolic Molecular Diagnosis of Inflammatory Bowel Disease by Synergistical Promotion of Layered Titania Nanosheets with Graphitized Carbon

Due to inefficient diagnostic methods, inflammatory bowel disease (IBD) normally progresses into severe complications including cancer. Highly efficient extraction and identification of metabolic fingerprints are of significance for disease surveillance. In this work, we synthesized a layered titania nanosheet doped with graphitized carbon (2D-GC-mTNS) through a simple one-step assembly process for assisting laser desorption ionization mass spectrometry (LDI-MS) for metabolite analysis. Based on the synergistic effect of graphitized carbon and mesoporous titania, 2D-GC-mTNS exhibits good extraction ability including high sensitivity (< 1 fmol µL^-1) and great repeatability toward metabolites. A total of 996 fingerprint spectra were collected from hundreds of native urine samples (including IBD patients and healthy controls), each of which contained 1220 m/z metabolite features. Diagnostic model was further established for precise discrimination of patients from healthy controls, with high area under the curve value of 0.972 and 0.981 toward discovery cohort and validation cohort, respectively. The 2D-GC-mTNS promotes LDI-MS to be close to clinical application, with rapid speed, minimum sample consumption and free of sample pretreatment.

Supplementary Information

The online version contains supplementary material available at 10.1007/s43657-022-00055-0.

Top-Down Rational Engineering of Heteroatom-Doped Graphene Quantum Dots for Laser Desorption/Ionization Mass Spectrometry Detection and Imaging of Small Biomolecules

Matrix-assisted laser desorption/ionization mass spectrometry imaging (MALDI MSI) is widely applied in mapping macrobiomolecules in tissues, but it is still limited in profiling low-molecular-weight (MW) compounds (typically metabolites) due to ion interference and suppression by organic matrices. Here, we present a versatile "top-down" strategy for rational engineering of carbon material-based matrices, by which heteroatom-doped graphene quantum dots (HGQDs) were manufactured for LDI MS detection and imaging of small biomolecules. The HGQDs derived from parent materials inherited the π-conjugated networks and doping sites for promoting energy transfer and negative ion generation, while their extremely small size guaranteed the matrix uniformity and signal reproducibility in LDI MSI. Compared to other HGQDs, nitrogen-doped graphene quantum dots (NGQDs) exhibited superior capability of assisting LDI of various small molecules, including amino acids, fatty acids, saccharides, small peptides, nucleobases, anticancer drugs, and bisphenol pollutants. Density functional theory simulations also corroborated that the LDI efficiency was markedly raised by the proton-capturing pyridinic nitrogen species and compromised by the electron-deficient boron dopants. NGQDs-assisted LDI MS further enabled label-free investigation on enzyme kinetics using an ordinary short peptide as the substrate. Moreover, due to the high salt tolerance and signal reproducibility, the proposed negative-ion NGQDs-assisted LDI MSI was able to reveal the abundance and distribution of low-MW species in rat brain tissue and achieved the imaging of low-MW lipids in coronally sectioned rat brains subjected to traumatic brain injury. Our work offers a new route for customizing nanomaterial matrices toward LDI MSI of small biomolecules in biomedical and pathological research.

NH2NH-MOF: a reaction matrix for the specific determination of small aldehydes by MALDI-MS

Efficient determination of aldehydes by matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) is hampered mainly by the low mass and unstable nature of analytes. In the present work, we propose a combined strategy of a reactive metal-organic framework (MOF) matrix for the derivatization and detection of aldehydes. A novel reactive MOF matrix (NH₂NH-MOF) was synthesized in two steps. First, NR₃⁺-MOF was synthesized via Cu²⁺ and the quaternary amine ligand 4,4'-bipyridinium, 1,1″-(1,2-ethanediyl)bis-, dibromide (PyEtBr). Then, -NHNH₂ was introduced to NR₃⁺-MOF through electrostatic adsorption between the -NR₃⁺ and -HSO₃^- of 4-hydrazinylbenzenesulfonic acid to synthesize NH₂NH-MOF. The acid-base chemistry of NH₂NH-MOF led to uniform cocrystallization of the aldehyde-matrix mixtures and helped to achieve the detection of low-weight aldehydes with good relative standard deviations (RSDs = 0.07-12.35%). It was confirmed that this strategy can accurately quantify formaldehyde, valeraldehyde, and benzaldehyde with good linearity (r > 0.97). Furthermore, this strategy was applied to quantitatively detect benzaldehyde in wastewater, thus showing potential applications in environmental pollutant detection.

Reactive Matrices for Analytical Matrix-Assisted Laser Desorption/Ionization (MALDI) Mass Spectrometry

In recent years, a special focus is placed on the usage of reactive matrices for analytical matrix assisted laser desorption/ionization mass spectrometry (MALDI-MS). Since 2003, when the term "reactive matrices" was suggested and the dignity of compounds, possessing dualistic properties as matrices and derivatization agents was demonstrated, corresponding approach has found application in various fields and, in particular, in bioanalysis (metabolomics, lipidomics, etc.). The main advantage of this methodology is that it reduces sample treatment time, simplifies the procedure of sample handling, improves the sensitivity of analysis, enhances the molecular identification and profiling. Within the framework of this review, the main attention is paid to "true" reactive matrices that interact with analyte molecules through an exchange or addition reactions. A special section discusses practical application of reactive matrices in the determination of the distribution of targeted and non-targeted organic substances on the surface of biological tissue sections by MALDI-MS imaging. In this critical review, a controversial proposal is made to consider protonating and deprotonating matrices as reactive, because they can undergo a chemical reaction such as proton transfer that occurs in both target solution and MALDI plume. In this respect, special attention is paid to "proton sponge" matrices that have found a wide application in the analysis of various acidic compounds by MALDI-MS in the negative mode. Historical data on the formation of ions and the fate of matrices in MALDI are considered at the beginning of this article.

Lignin as a MALDI matrix for small molecules: a proof of concept

Driven by the interest in metabolomic studies and the progress of imaging techniques, small molecule analysis is booming, while it remains challenging to be realized by matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS). Herein, lignin, the second most abundant biomass in nature, was applied as a dual-ion-mode MALDI matrix for the first time to analyze small molecules. The low ionization efficiency and strong optical absorption properties make lignin a potential MALDI matrix in small molecule analysis. A total of 30 different small molecules were identified qualitatively and six kinds of representative molecules were detected quantitatively with a good linear response (R² > 0.995). To verify the accuracy of our quantitative method in MALDI, myricitrin, a major bioactive component in Chinese bayberry, was analyzed in different cultivars and tissues. The myricitrin content in real samples detected by MALDI was highly consistent (R² > 0.999) with that detected by high-performance liquid chromatography, thus indicating the applicability of the lignin matrix. Further characterization by ultraviolet and nuclear magnetic resonance spectroscopy was carried out to explain the possible mechanism of lignin as a matrix and provide more theories for a rational matrix design.

A dual-mode reactive matrix for sensitive and quantitative analysis of carbohydrates by MALDI-TOF MS

Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) is a widely used tool for the analysis of carbohydrates. However, the detection of carbohydrates by MALDI-TOF MS is often limited by the unsatisfactory ionization efficiency, instability and the matrix interference in low molecular weight region. Here, we demonstrate that a reactive matrix, 2-hydrazinoquinoline (2-HQ), can be used to detect neutral, sialic and low molecular weight carbohydrates sensitively both in the positive and negative ion mode. Since 2-HQ reacts efficiently with the reducing end of carbohydrate to form stable hydrazone, the ionization efficiency of derived carbohydrates is significantly enhanced. Using 2-HQ, the sensitivity for analyzing glycans has been improved 10-fold and 100-fold compared with those using 3-aminquinoline (3-AQ) and 2,5-dihydroxybenzoic acid (DHB) as matrix, respectively. Moreover, quantitative analysis of neutral, acidic and low molecular weight carbohydrates has been achieved because of the good reproducibility by using 2-HQ as matrix. As a result, up to 50 glycans in a single sample spot of human fresh serum without any prior purification and enrichment have been successfully detected. Therefore, 2-HQ as a new reactive matrix has shown great potentials in widespread applications for sensitive, selective, quantitative, high speed and high throughput analysis of carbohydrates in complex samples by MALDI-TOF MS.

4-Hydrazinoquinazoline acting as a reactive matrix for the rapid and sensitive analysis of neutral and sialylated glycans using MALDI MS

The direct analysis of glycans by matrix-assisted laser desorption/ionization mass spectrometry (MALDI MS) presents limited sensitivity due to the lower ionization efficiency of glycans. Various chemical derivatization methods have been developed to improve the detection sensitivity of glycans, but most of them need tedious preparation and cleanup procedures. Herein, a reactive matrix, 4-hydrazinoquinazoline (4-HQ), was developed for the rapid and sensitive detection of both neutral and sialylated glycans by MALDI MS. With 4-HQ as the reactive matrix, the detection limits of maltoheptaose and A3 glycan decreased 100-fold and 20-fold, respectively, compared with the conventional matrix. Moreover, 4-HQ formed homogeneous crystals and therefore showed good shot-to-shot reproducibility. Finally, the reactive matrix was successfully applied for the analysis of glycans released from glycoproteins and human serum. Importantly, the application of 4-HQ is the same as that of a conventional matrix with the additional advantage of on-target reaction at room temperature. Thus, 4-HQ can be used for the routine analysis of glycans by MALDI MS due to its simple use, great reproducibility, and enhanced detection of both neutral and sialylated glycans.

Advances in mass spectrometry-based glycomics-An update covering the period 2017-2021

The wide variety of chemical properties and biological functions found in proteins is attained via post-translational modifications like glycosylation. Covalently bonded to proteins, glycans play a critical role in cell activity. Complex structures with microheterogeneity, the glycan structures that are associated with proteins are difficult to analyze comprehensively. Recent advances in sample preparation methods, separation techniques, and MS have facilitated the quantitation and structural elucidation of glycans. This review focuses on highlighting advances in MS-based techniques for glycomic analysis that occurred over the last 5 years (2017-2021) as an update to the previous review on the subject. The topics of discussion will include progress in glycomic workflow such as glycan release, purification, derivatization, and separation as well as the topics of ionization, tandem MS, and separation techniques that can be coupled with MS. Additionally, bioinformatics tools used for the analysis of glycans will be described.

Core-shell MOF@COFs used as an adsorbent and matrix for the detection of nonsteroidal anti-inflammatory drugs by MALDI-TOF MS

A core-shell material (UiO@TapbTp) has been developed as an adsorbent and matrix to detect nonsteroidal anti-inflammatory drugs (NSAIDS) by matrix laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) in complex samples. The hybrid material is prepared by growing covalent organic framework (COF, TapbTp) layers in situ on an amino-modified metal-organic framework (MOF, UiO-66-NH₂). The combination of the MOF and COF overcomes their individual shortcomings and integrates both of their advantages. Compared with the bare COF and MOF, the core-shell composite exhibits improved enrichment ability and matrix performance. With the help of pre-enrichment under optimized conditions, the limits of detection (LODs) for ketoprofen, naproxen, and aspirin are reduced by nearly 1000 times, with values of 0.001 mg L^-1, 0.010 mg L^-1, and 0.001 mg L^-1, respectively, and the relative standard deviations (RSDs) are all below 12.35%. The good recoveries (84.8-118%) in (spiked) saliva and environmental water sample further verify the applicability of the method in complex samples.

Self-assembled matrix fabricated by Fe-metal organic frameworks and carboxymethyl cellulose for the determination of small molecules by MALDI-TOF MS

A nanoprobe of laser desorption/ionization-time of flight mass spectrometry (LDI-TOF MS) for the determination of small molecules was developed that is based on the composition of Fe-metal organic frameworks (Fe-MOFs) and carboxymethyl cellulose-Na (CMC-Na). This material is a good adsorbent for small molecules via hydrogen bonding and π-interactions; we detected three molecules, dopamine, glyphosate, and pyrene. The detection limits for these compounds are 0.01 mg L^-1, 1.50 μg L^-1, and 0.01 μg L^-1, respectively; the recoveries are 85-117%, 81-127%, and 89-115%, respectively. The relative standard deviations (~ 15%) and coefficients of determination of the calibration plot (~ 0.97) are satisfactory. The applicability of the chip for practical samples is demonstrated by quantifying pyrene in domestic water and polluted lake water; the recoveries are about 90~117% and 85~125% (n = 5), respectively; the RSDs are 9.4% and 13.5%, respectively. Graphical abstract.

Parallel On-Target Derivatization for Mass Calibration and Rapid Profiling of N-Glycans by MALDI-TOF MS

Glycosylation is an important post-translational modification of proteins, and abnormal glycosylation is involved in a variety of diseases. Accurate and rapid profiling of N-glycans by matrix-assisted laser desorption ionization-time-of-flight mass spectrometry (MALDI-TOF MS) is still technically challenging and hampered mainly by mass drift of instrument, manual identification of spectrum peaks, and poor cocrystallization with traditional matrices besides low ionization efficiency of analytes. In the present study, a parallel on-target derivatization strategy (POTDS), on the basis of two rationally combined matrices, i.e., 3-hydrazinobenzoic acid plus DHB (DHB/3HBA) and quinoline-3-carbohydrazide plus DHB (DHB/Q3CH), was proposed for mass calibration and rapid detection of reducing N-glycans. Both DHB/3HBA and DHB/Q3CH show high derivatization efficiency and can improve the ionization efficiency of reducing N-glycans significantly. For mass calibration, in combination with dextrans, DHB/3HBA and DHB/Q3CH prove to be highly sensitive matrices facilitating both MS and MS² calibration for N-glycans in dual polarities. For rapid identification, the regular mass difference observed for each N-glycan labeled with Q3CH and 3HBA respectively can eliminate the occurrence of false positives and promote automated identification of N-glycans in complex samples. For relative quantitation, the acid-base pair of DHB/Q3CH generates a concentrated cocrystallization of glycan-matrix mixtures at the edge of the droplet uniformly, exhibiting good linearity (R² > 0.998) and accuracy (RSD ≤ 10%). Furthermore, the established POTDS was successfully utilized to assess N-glycans of serum from HCC patients, revealing potential for biomarker discovery in clinical practice.