A narrative review of metabolomics in the era of “-omics”: integration into clinical practice for inborn errors of metabolism

Introduction

Metabolites are small organic molecules (≤1,000 m/z) that are the intermediate or end products of enzymatic processes. Their levels in cells, tissues, or biological fluids may vary due to gene function, disease processes, diet, environment, medications, and other factors (1,2).

Inborn errors of metabolism (IEMs), although individually rare, are a key group of over 1,000 inherited disorders that result in altered levels of metabolites (3). Some of these disorders affect the processing of carbohydrates, lipids, or proteins, leading to intoxication or energy deficiency in the case of small molecules, and storage or deficiency in the case of complex molecules (4). In many of these conditions, early diagnosis allows for timely medical intervention, reducing mortality and significantly improving quality of life. Such prompt diagnoses of IEMs can be achieved through newborn screening (NBS) programs, which involve dried blood spot (DBS) analysis using mass spectrometry techniques. Depending on the country and region, these programs can detect around 30 disorders in the pre-symptomatic phase, facilitating early intervention and improving clinical outcomes (5-8).

Diagnosis of many IEMs, either through NBS or diagnostic testing, requires metabolomic analysis. This traditional approach to the investigation of metabolites is referred to as “targeted metabolomics”. Different techniques (such as gas or liquid chromatography-mass spectrometry) are used to analyze a predefined, small list of metabolites so that they can be separated, detected, annotated, and quantified. For decades, these approaches have been widely and successfully used for the diagnosis and monitoring of patients with IEMs.

More recently, a powerful “next-generation metabolic screening” technology has become available to assess the metabolome in an unbiased fashion. This has been facilitated by the development of high-resolution accurate mass spectrometers, as well as progress in the understanding of biochemical pathways and bioinformatics. As a result, untargeted metabolomics has significantly matured in the last decade and offers an increased diagnostic yield when compared to traditional approaches (9-14). The applications of both targeted and untargeted metabolomics may be varied; however, one crucial application is in the diagnosis and therapeutic monitoring of patients with IEMs (15,16).

Similar to what has occurred in the field of genomic testing with next generation sequencing (NGS), untargeted metabolomics is beginning to alter the diagnostic approach to the investigation of IEMs due to its broader coverage and scope. Over the past two decades untargeted metabolomic approaches are increasingly being utilized in the investigation and diagnosis of patients with IEMs (15).

Not only does untargeted metabolomics assist with diagnosis, it also allows for increased understanding of disease mechanisms (17-19), customization of drug treatments (20), and monitoring of therapeutic response (21-24). As genomic studies unravel new disorders, understanding their effects on cellular and subcellular function becomes essential to confirm causation of disease and develop effective therapies (25). In addition to the use of traditional sample types (urine, plasma and cerebrospinal fluid), other tissue and cell types are also amenable to metabolomic studies, which will advance the understanding of intracellular and organ-specific metabolic pathways (26). Concurrently, the other rapidly expanding “-omic” fields (including transcriptomics, epigenomics, proteomics, lipidomics and glycomics) will also be increasingly contributing to these improvements (25,27).

This review aims to highlight the clinical utility of untargeted metabolomics, and its complementary nature to currently utilized “-omic” approaches, for the diagnosis and monitoring of IEMs. Current methods and limitations of untargeted metabolomics will be reviewed, in comparison to targeted metabolomics. Future applications of untargeted metabolomics in the field of IEMs, including identifying ‘metabolomic signatures’, monitoring therapeutic responses and integrating genomic and metabolomic data, will also be discussed. Over time, widespread availability of this technology will improve understanding of molecular mechanisms of disease and increase diagnostic yield, resulting in positive impacts on patient outcomes. We present the following article in accordance with the Narrative Review reporting checklist (available at https://tp.amegroups.com/article/view/10.21037/tp-22-105/rc).

Methods

In this narrative review, authors performed a literature search using PubMed and Google Scholar using the MeSH terms: “Metabolomic”, “Diagnosis”, “Untargeted Metabolomics”, “Inborn Errors of Metabolism”. Article inclusion and exclusion criteria are outlined in Table 1. Additionally, where applicable, authors also utilized their own repositories of articles.

Evaluating the metabolome

For targeted metabolomics, gas chromatography-mass spectrometry and liquid chromatography tandem mass spectrometry (LC/MS-MS) are both heavily utilized in the clinical setting, as sample preparation is often simple and analytical results can be available rapidly. With these techniques, a finite number of compounds are analyzed, allowing optimization of sensitivity and specificity (28,29). There are now universal databases that provide reference spectra (30).

In untargeted metabolomics, the sheer number and complexity of metabolites detected is far greater (even up to 40,000 metabolites has been reported) (14) and provides an unbiased view of the entire metabolome similar to whole genome sequencing (WGS) in genomic evaluation. Untargeted metabolomics can be performed by ultra-high performance mass spectrometry based approaches in tandem with liquid chromatography (31,32). Additionally, there are other modern technologies and methodologies that are being employed to detect large numbers of metabolites, including the ultra-high resolution of Orbitrap-mass spectrometry which leads to accurate detection of molecular ions (<1 ppm mass deviations) with extremely high specificity (33). Matrix-associated laser desorption ionization time-of-flight (MALDI-TOF) mass spectrometry carries the benefits of short sample analysis time, wide mass range of detectable metabolites and low sample consumption (<1 µL) (34). Other platforms, such as ¹H-nuclear magnetic resonance (NMR) spectral analysis, are now employed in both targeted and untargeted approaches (35,36). However, this mini-review will focus primarily on mass spectrometry-based techniques.

Automated bioinformatic curation, based on population z-scores, reduces the number of compounds requiring manual review downstream (37,38). Similar to genomic pipelines assisting in interpretation of large volumes of genomic data, biochemical pipelines that help streamline and prioritize metabolite analysis have also been developed (10).

Current diagnostic pathways: targeted metabolomics

Traditionally, IEMs have been diagnosed in a targeted manner (driven by phenotypic or molecular findings) (29,39). Such biomarker assays, as outlined by Saudubray and colleagues, are developed and validated with a high level of sensitivity and specificity for disease and are suitable for use in clinical practice (4). This targeted approach streamlines the diagnostic pathway, allowing for timely implementation of treatment in patients presenting during an acute metabolic crisis. However, a targeted approach comes with limitations in terms of the quantity of metabolites analyzed, as well as an inability to easily assess interactions between metabolic pathways.

With the targeted approach, patients presenting with clinical features of an IEM (acute metabolic encephalopathy, with basic biochemistry suggestive of small molecule intoxication), are rapidly diagnosed through tests such as urine metabolic screening (including organic acids), plasma acylcarnitine and/or amino acid analyses (29). Other patients presenting with a more chronic process due to accumulation of a complex molecule, such as glycosaminoglycans, sphingolipids or very long chain fatty acids, have a separate set of metabolomic biomarkers that are evaluated to arrive at a diagnosis. The benefits of a targeted metabolomic approach include comprehensive understanding of the relevant enzyme pathways, optimized sample preparation (with reduction in high-abundance molecules) and the ability to more effectively filter out analytical artefacts (40,41). Newer technologies, such as flow injection mass spectrometry, can reduce run times, enabling faster analysis that could be suitable for high-throughput testing (38).

With the increasing utility of NGS technologies, targeted metabolomic studies are also playing a role in assisting with molecular diagnostics. Detection of a specific metabolic derangement can facilitate a simpler bioinformatic gene panel analysis, rather than an open whole exome approach with a non-specific phenotype (42). However, it is important to consider that the absence of a metabolomic marker does not preclude a diagnosis, as these may only be elevated during times of metabolic stress.

Whilst the clinical utility of metabolomics currently lies in targeted approaches, advances in technology and bioinformatics have facilitated untargeted metabolic analysis in both the clinical and research spheres, allowing for subsequent integration into the clinical workflow.

Future trends and integration of untargeted metabolomics to the bedside

Untargeted metabolomic technologies allow for analysis of a comprehensive array of metabolites that would otherwise require individualized assays and sample preparations (43). This technology is also likely to offer substantial clinical benefit not achieved by targeted metabolomic testing, and may be further enhanced by developments in machine learning (44) and statistical modelling (45,46). The clinical potential of untargeted metabolomics in diagnosis of IEM has been investigated extensively (11,13,15,18,22,33,47-53), and complements other “-omic” approaches in the diagnosis of IEM and other genetic disorders (16,42,54-58).

Biomarker discovery and increased understanding of disease pathophysiology

A key application of untargeted metabolomics is in identification of novel, or more sensitive biomarkers for disease. To be considered an ideal biomarker (for diagnostic and prognostic features) of disease, the candidate analytes derived in a research setting need to be validated extensively before translating into clinical practice (59). There are multiple considerations when identifying an ideal disease biomarker. These analytes must be sensitive, in that patients with the disease should have a derangement in the biomarker (such as allo-isoleucine in maple syrup urine disease) (60). They should be specific; for instance, lactate is a poor biomarker for IEMs because its elevation can occur due to multiple other aetiologies (including sepsis, seizures or cardiac dysfunction) (61). The extent of biomarker derangement should correlate with disease severity, as well as clinical response to therapy. There are also analytical factors that should be considered, such as sample robustness (during handling, storage and freeze-thaw cycles) and reproducibility (62,63). It is unlikely that new candidate biomarkers would be found using targeted, or even semi-targeted approaches, given the limited number of metabolites and pathways interrogated. However, with untargeted analytical approaches, the unbiased screening of the metabolome facilitates biomarker discovery.

For instance, Burrage and colleagues (17) performed untargeted metabolomics using mass spectrometry on plasma samples of 48 patients with urea cycle disorders. This involved the detection of over 900 different metabolites and normalization of subject samples to an invariant anchor specimen, which could then be compared to population reference ranges to generate z-scores. Metabolites of sodium benzoate and sodium phenylbutyrate were detected, allowing for monitoring of therapeutic compliance. In addition, marked increases in multiple guanidino compounds were observed in patients with arginase deficiency, which were elevated over-and-above the plasma arginine level. Wangler and colleagues studied individuals with confirmed peroxisome biogenesis disorders in the Zellweger spectrum (PBD-ZSD). The untargeted metabolomic profiling results from plasma detected over 650 compounds, with increases observed in pipecolic acid and long-chain lysophosphatidylcholines, along with unexpected decreases in multiple sphingomyelin species (64). These findings could serve as more sensitive biomarkers for the initial diagnosis, act as early signs of impending clinical deterioration, determine prognosis or prove beneficial in monitoring response to novel therapies. Once identified, these biomarkers, or patterns of biomarkers can be incorporated into targeted metabolomic analysis.

Identification of novel markers of disease also provides additional insight into disease mechanisms and pathogenesis. Glinton and colleagues, utilizing untargeted metabolomic profiling for four patients with serine biosynthetic defects, demonstrated low levels of glycerophosphocholine, glycerophosphoethanolamine and sphingomyelin, as well as the well-known deficiencies of serine and glycine (18). Most of these phospholipid compounds normalized following treatment with serine and glycine. The authors suggested that deficiency of phospholipids (secondary to serine deficiency) may be contributing to the neurological manifestations of the condition, and this could lend itself to additional targeted therapies in the future. As exemplified by this study, and others, untargeted methods allow for a better understanding of disease pathogenesis, as whole pathways can be analyzed, rather than individual metabolites (32,56,57,65-69).

The use of untargeted approaches can also help generate a ‘metabolomic signature’ similar to the increasing utility of methylation episignatures for detection of BAFopathies (70). Venter and colleagues performed untargeted metabolomic testing on urine samples for patients with suspected respiratory chain disorders, identifying a 12-compound signature that distinguished patients with biopsy-proven respiratory chain abnormalities in muscle, from those with a suspicion of a mitochondrial myopathy who did not have these changes (with a sensitivity of 98%, and specificity of 80%) (71). There are also several other reports of utilization of untargeted metabolomics to generate disease specific ‘fingerprints’ (58,72-74).

Additionally, diagnostic evaluation of samples from 16 patients with pyruvate kinase deficiency (compared to 32 controls) generated a highly specific and accurate metabolomic signature that included glycolytic intermediates, polyamines and acylcarnitine species (75). Such comparison of patient samples to a library of known metabolomic signatures could be beneficial when a more targeted approach has not achieved a diagnosis (Figure 1). Similar to the 100,000 Genomes Pilot Study (76), cohort analysis of large numbers of such patients may also help identify and refine candidate biomarkers and metabolomic signatures.

Figure 1 The integration of untargeted metabolomics into the diagnostic pipeline of IEMs. Patients who have no diagnosis despite targeted metabolomic and molecular testing are candidates for an untargeted approach, including metabolomics and WGS. Novel biomarkers discovered during this process can then be re-incorporated into targeted metabolomic analysis. Any VUSs unresolved by the combined genomic-metabolomic approach can be further investigated through functional characterization studies in the research setting. VUS, variant of uncertain significance; IEMs, inborn errors of metabolism; WGS, whole genome sequencing.

Using ‘metabolomic signatures’ to identify candidate biomarkers for treatment response

With increased throughput and understanding of individual metabolites in disease pathogenesis, metabolomic signatures could be built upon to incorporate response to treatment. There are already examples in the literature of untargeted metabolomic analysis assisting in treatment monitoring for IEMs (77,78). Pillai and colleagues reported a patient with riboflavin transporter deficiency with bi-allelic variants one pathogenic and one a variant of uncertain significance (VUS) in SLC52A2 (77). Untargeted metabolomic analysis identified multiple compounds associated with abnormal flavin adenine nucleotide function, which normalized after riboflavin therapy. This both assisted in confirmation of the molecular diagnosis and provided a mechanism for identification of candidate biomarkers that can potentially be validated to be used for monitoring therapeutic response through targeted assays.

These cases demonstrate that, aside from its diagnostic role, untargeted metabolomics may also identify biomarkers that assist in therapeutic decision-making. This not only applies to patients with IEMs, but also in other diseases with larger health economic considerations, including cancer, cardiovascular disease and infectious diseases (79). Untargeted metabolomics may also facilitate a personalized medical approach that is not only patient-specific, but also incorporates contemporaneous external factors such as diet, treatment, and environment. Although outside the scope of this review, the recent advances in single-cell (or organelle) metabolomics also have the potential to revolutionize the treatment approaches in cancer as well as assist in understanding of cell senescence (80).

Incorporation of untargeted metabolomics into NBS

Given the comprehensive analysis provided by untargeted metabolomics, it may play a significant role in the future of NBS. Untargeted metabolomics can be successfully performed on DBS (81-83). To our knowledge, results of a large-scale study involving untargeted metabolomics in an asymptomatic population are not available. Liu and colleagues reviewed 4,464 clinical samples analyzed via targeted metabolomics, compared with 2,000 samples (from a separate cohort) analyzed using untargeted techniques (12). The former had a 1.3% diagnostic rate [14 conditions identified, three of which were not in the Recommended Uniform Screening Panel (RUSP) (84)]. The latter had a diagnostic rate of 7.1%, but this included 49 conditions that were not included in the RUSP (19 of which do not have any known treatment). This study demonstrates that untargeted metabolomics can identify conditions detected by NBS; however, the risks of identifying additional disorders not routinely included in the RUSP are also high. Typically, the RUSP scoring system considers diagnostic and clinical evidence for a disease (and its treatment) prior to recommendation: identifying non-RUSP conditions, some of which are less defined clinical entities, may create more uncertainty.

Over a decade ago, the National Institutes of Health (NIH)-funded Newborn Sequencing In Genomic medicine and public HealTh (NSIGHT) Consortium commenced pilot programs for whole exome sequencing (WES)- and WGS-based NBS in the United States of America (85). Many lessons have been learned from these initiatives including those associated with technical issues (data analysis and handling) and those associated with ethical and societal challenges (secondary findings of adult-onset disorders for which no therapies are currently available) (86). The latter pose a significant ethical dilemma with the possibility of interfering with the neonate’s right to self-determination (85). A NBS pilot study of a combined genomic-untargeted metabolomic approach, which only analyzes genes and pathways associated with treatable childhood-onset conditions, may help to reduce this risk of secondary findings.

Untargeted metabolomics as a diagnostic companion to genomics

With more rapid analytical techniques (87) now available, untargeted metabolomic testing is complementing genomic approaches for the diagnosis of patients suspected of having an IEM. For example, in a patient with intellectual disability and skeletal dysplasia, untargeted metabolomics identified the unusual metabolite N-acetylated mannosamine. As a result, two missense variants in NANS detected by WGS were prioritized for further work-up (88). This case was combined with eight other similar patients ascertained through WGS, and further functional work confirmed impairment in NANS enzyme activity as the basis for these patients’ clinical condition.

In other instances where molecular testing has already identified VUSs, untargeted metabolomic analysis can contribute significantly to variant interpretation (42,55,89-92). In their “cross-omics” study, Kerkhofs and colleagues showed that for accurate prioritization of disease-causing genes in IEMs, it is essential to take into account the primary pathway of the affected protein as well as the broader network of metabolites (55). Alaimo and colleagues reviewed data on 170 patients who had both WES and untargeted metabolomics (89). The metabolomic data contributed to variant interpretation in 74 individuals (43.5%) and confirmed a clinical diagnosis in 21 of these cases (12.3% diagnostic rate). The American College of Medical Genetics and Genomics (ACMG) criteria give specific provision for the incorporation of results generated by such high-quality functional validation studies into variant classification (93,94). Increasingly, the combination of “-omic” technologies will provide the best approach for re-evaluating previously unsolved cases.

Limitations of untargeted metabolomics

As with all analytical techniques, there are significant limitations and challenges with untargeted metabolomics (95). Similar to targeted metabolomics, the instrumentation required for untargeted metabolomics is expensive to both purchase and maintain (96). The specialized methodology requires a high level of expertise across both wet lab sample preparation through to dry lab results interpretation (97). Additionally, when compared with other “-omic” technologies (such as genomics), metabolomic results are heavily influenced by physiological factors such as gender, diet, and drug treatment. Age is also a particularly important confounder, as metabolomic signatures are influenced by both gestational age and chronological age (98,99). However, newer regression models are being developed to control for these factors (100). Additionally, ambient temperature and humidity can affect sample preparation (101) and assay implementation steps which in turn can influence the data produced and associated interpretation of metabolic profiling.

The bioinformatic pipelines that are required for untargeted metabolomic analysis also have specific challenges. For instance, bioinformatic pre-processing methods can impact results under different settings (102). Metabolite identification (including consideration of the associated pathways) is both complex and time consuming (2). Also, the specialized expertise required for analysis and interpretation needs to be integrated into the laboratory workflow for fast and accurate clinical outcomes (103).

Drawing parallels with genomics, early allele frequency databases lacked ethnic diversity, which negatively impacted variant curation in minority groups (104,105). Similarly, more work is required to compile the “metabolomes” of different populations and ethnic groups around the world for metabolomics to address ethnic specificity of a given disease (106). One key additional challenge will be “filtering out” metabolite variation seen in normal populations during catabolic stress (107), to avoid misattribution of normal physiology to an IEM. As point of difference from genomic investigations, reference libraries for metabolite identification are not as well established as genomic reference sequences. Additionally, fragmentation and isotope patterns, m/z ratios, and retention time data are available for many yet unidentified metabolites; however, this is changing with increased testing. For instance, the most recent update to the Human Metabolome Database (HMDB) includes a nearly two-fold increase in the number of annotated metabolites (from 114,000 to 217,920), as well as improved spectral and pathway visualization tools (108).

Many IEMs have organ-specific manifestations, such as encephalopathy and seizures, or isolated cardiomyopathy (109,110). These conditions are likely to have distinct metabolomic profiles in different tissues and sample types, which also need to be accounted for in the analysis.

Given the above limitations, the sensitivity of untargeted metabolomics for diagnosing established IEMs is lower than that of targeted metabolomics. Almontashiri and colleagues performed untargeted metabolomic analysis in 87 patients with confirmed IEMs, with a sensitivity of 86% (111). This is analogous to the challenges with coverage observed in WES and WGS, as compared to targeted testing (112). Steinbusch and colleagues reported similar sensitivity (87%), with seven out of the 10 ‘missed’ cases being patients already on therapy (113).

Many of the other difficulties are likely to be overcome with increasing use of untargeted metabolomics. Testing will become more affordable over time, resulting in increased accessibility. Normal population databases (including ethnically diverse minority populations) and bioinformatic pipelines will become more robust, facilitating identification of pathological states. Considering the incredible potential of untargeted metabolomics in revolutionizing the diagnosis, monitoring and treatment of IEMs, the challenges are outweighed by the benefits.

Conclusions

Metabolomics originally developed as a targeted testing methodology, quantifying a panel of key metabolites to diagnose a limited number of clinically relevant IEMs. However, with the recent advances in technology, untargeted metabolomics is increasingly being applied in screening for a broad range of rare diseases, as well as for functional validation of genomic variants, identification of drug targets, and monitoring response to therapy. Decreasing costs over time will allow for increased use of untargeted metabolomics, in the field of IEMs but also in others such as oncology, cardiology and infectious diseases. The more holistic analytical approach will considerably improve understanding of the pathophysiology of many IEMs, as well as identifying additional secondary metabolites to create disease-specific metabolic signatures. Tissue-specific, single cell, and even single organelle, metabolomic investigations also offer the opportunity to diagnose disorders with selective expression and can help improve treatment options for many diseases. The potential applications in screening of asymptomatic populations (as in NBS) may have significant health economic benefits: however, it would be essential to reduce the risk of secondary findings by using appropriate bioinformatic curation tools. Integration of untargeted metabolomics, along with genomics, into routine clinical care is likely to improve diagnostic workflows and patient management, resulting in both improved clinical and health economic outcomes.

Acknowledgments

Funding: None.

Footnote

Reporting Checklist: The authors have completed the Narrative Review reporting checklist. Available at https://tp.amegroups.com/article/view/10.21037/tp-22-105/rc

Peer Review File: Available at https://tp.amegroups.com/article/view/10.21037/tp-22-105/prf

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tp.amegroups.com/article/view/10.21037/tp-22-105/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Klassen A, Faccio AT, Canuto GA, et al. Metabolomics: Definitions and Significance in Systems Biology. Adv Exp Med Biol 2017;965:3-17. [Crossref] [PubMed]
2. Tebani A, Abily-Donval L, Afonso C, et al. Clinical Metabolomics: The New Metabolic Window for Inborn Errors of Metabolism Investigations in the Post-Genomic Era. Int J Mol Sci 2016;17:1167. [Crossref] [PubMed]
3. Ferreira CR, van Karnebeek CDM. Inborn errors of metabolism. Handb Clin Neurol 2019;162:449-81. [Crossref] [PubMed]
4. Saudubray JM, Mochel F, Lamari F, et al. Proposal for a simplified classification of IMD based on a pathophysiological approach: A practical guide for clinicians. J Inherit Metab Dis 2019;42:706-27. [Crossref] [PubMed]
5. Chace DH, Kalas TA, Naylor EW. Use of tandem mass spectrometry for multianalyte screening of dried blood specimens from newborns. Clin Chem 2003;49:1797-817. [Crossref] [PubMed]
6. Heringer J, Valayannopoulos V, Lund AM, et al. Impact of age at onset and newborn screening on outcome in organic acidurias. J Inherit Metab Dis 2016;39:341-53. [Crossref] [PubMed]
7. Wilcken B. Screening for disease in the newborn: the evidence base for blood-spot screening. Pathology 2012;44:73-9. [Crossref] [PubMed]
8. Wilcken B, Wiley V. Fifty years of newborn screening. J Paediatr Child Health 2015;51:103-7. [Crossref] [PubMed]
9. Coene KLM, Kluijtmans LAJ, van der Heeft E, et al. Next-generation metabolic screening: targeted and untargeted metabolomics for the diagnosis of inborn errors of metabolism in individual patients. J Inherit Metab Dis 2018;41:337-53. [Crossref] [PubMed]
10. Hoegen B, Zammit A, Gerritsen A, et al. Metabolomics-Based Screening of Inborn Errors of Metabolism: Enhancing Clinical Application with a Robust Computational Pipeline. Metabolites 2021;11:568. [Crossref] [PubMed]
11. Kennedy AD, Miller MJ, Beebe K, et al. Metabolomic Profiling of Human Urine as a Screen for Multiple Inborn Errors of Metabolism. Genet Test Mol Biomarkers 2016;20:485-95. [Crossref] [PubMed]
12. Liu N, Xiao J, Gijavanekar C, et al. Comparison of Untargeted Metabolomic Profiling vs Traditional Metabolic Screening to Identify Inborn Errors of Metabolism. JAMA Netw Open 2021;4:e2114155. [Crossref] [PubMed]
13. Miller MJ, Kennedy AD, Eckhart AD, et al. Untargeted metabolomic analysis for the clinical screening of inborn errors of metabolism. J Inherit Metab Dis 2015;38:1029-39. [Crossref] [PubMed]
14. Schrimpe-Rutledge AC, Codreanu SG, Sherrod SD, et al. Untargeted Metabolomics Strategies-Challenges and Emerging Directions. J Am Soc Mass Spectrom 2016;27:1897-905. [Crossref] [PubMed]
15. Kennedy AD, Wittmann BM, Evans AM, et al. Metabolomics in the clinic: A review of the shared and unique features of untargeted metabolomics for clinical research and clinical testing. J Mass Spectrom 2018;53:1143-54. [Crossref] [PubMed]
16. Mordaunt D, Cox D, Fuller M. Metabolomics to Improve the Diagnostic Efficiency of Inborn Errors of Metabolism. Int J Mol Sci 2020;21:1195. [Crossref] [PubMed]
17. Burrage LC, Thistlethwaite L, Stroup BM, et al. Untargeted metabolomic profiling reveals multiple pathway perturbations and new clinical biomarkers in urea cycle disorders. Genet Med 2019;21:1977-86. [Crossref] [PubMed]
18. Glinton KE, Benke PJ, Lines MA, et al. Disturbed phospholipid metabolism in serine biosynthesis defects revealed by metabolomic profiling. Mol Genet Metab 2018;123:309-16. [Crossref] [PubMed]
19. Tebani A, Abily-Donval L, Schmitz-Afonso I, et al. Unveiling metabolic remodeling in mucopolysaccharidosis type III through integrative metabolomics and pathway analysis. J Transl Med 2018;16:248. [Crossref] [PubMed]
20. Sahebekhtiari N, Nielsen CB, Johannsen M, et al. Untargeted Metabolomics Analysis Reveals a Link between ETHE1-Mediated Disruptive Redox State and Altered Metabolic Regulation. J Proteome Res 2016;15:1630-8. [Crossref] [PubMed]
21. Cappuccio G, Pinelli M, Alagia M, et al. Biochemical phenotyping unravels novel metabolic abnormalities and potential biomarkers associated with treatment of GLUT1 deficiency with ketogenic diet. PLoS One 2017;12:e0184022. [Crossref] [PubMed]
22. Pappan KL, Kennedy AD, Magoulas PL, et al. Clinical Metabolomics to Segregate Aromatic Amino Acid Decarboxylase Deficiency From Drug-Induced Metabolite Elevations. Pediatr Neurol 2017;75:66-72. [Crossref] [PubMed]
23. Gertsman I, Gangoiti JA, Nyhan WL, et al. Perturbations of tyrosine metabolism promote the indolepyruvate pathway via tryptophan in host and microbiome. Mol Genet Metab 2015;114:431-7. [Crossref] [PubMed]
24. Tallis E, Karsenty CL, Grimes AB, et al. Untargeted metabolomic profiling in a patient with glycogen storage disease Ib receiving empagliflozin treatment. JIMD Rep 2022;63:309-15. [Crossref] [PubMed]
25. Wevers RA, Blau N. Think big - think omics. J Inherit Metab Dis 2018;41:281-3. [Crossref] [PubMed]
26. Saoi M, Britz-McKibbin P. New Advances in Tissue Metabolomics: A Review. Metabolites 2021;11:672. [Crossref] [PubMed]
27. Sadikovic B, Levy MA, Kerkhof J, et al. Clinical epigenomics: genome-wide DNA methylation analysis for the diagnosis of Mendelian disorders. Genet Med 2021;23:1065-74. [Crossref] [PubMed]
28. Peterson AC, Russell JD, Bailey DJ, et al. Parallel reaction monitoring for high resolution and high mass accuracy quantitative, targeted proteomics. Mol Cell Proteomics 2012;11:1475-88. [Crossref] [PubMed]
29. Pitt JJ, Eggington M, Kahler SG. Comprehensive screening of urine samples for inborn errors of metabolism by electrospray tandem mass spectrometry. Clin Chem 2002;48:1970-80. [Crossref] [PubMed]
30. Ludwig C, Easton JM, Lodi A, et al. Birmingham Metabolite Library: a publicly accessible database of 1-D 1H and 2-D 1H J-resolved NMR spectra of authentic metabolite standards (BML-NMR). Metabolomics 2012;8:8-18. [Crossref]
31. Evans AM, DeHaven CD, Barrett T, et al. Integrated, nontargeted ultrahigh performance liquid chromatography/electrospray ionization tandem mass spectrometry platform for the identification and relative quantification of the small-molecule complement of biological systems. Anal Chem 2009;81:6656-67. [Crossref] [PubMed]
32. Tebani A, Abily-Donval L, Schmitz-Afonso I, et al. Analysis of Mucopolysaccharidosis Type VI through Integrative Functional Metabolomics. Int J Mol Sci 2019;20:446. [Crossref] [PubMed]
33. Bonte R, Bongaerts M, Demirdas S, et al. Untargeted Metabolomics-Based Screening Method for Inborn Errors of Metabolism using Semi-Automatic Sample Preparation with an UHPLC- Orbitrap-MS Platform. Metabolites 2019;9:289. [Crossref] [PubMed]
34. Ren JL, Zhang AH, Kong L, et al. Advances in mass spectrometry-based metabolomics for investigation of metabolites. RSC Adv 2018;8:22335-50. [Crossref] [PubMed]
35. Bingol K. Recent Advances in Targeted and Untargeted Metabolomics by NMR and MS/NMR Methods. High Throughput 2018;7:9. [Crossref] [PubMed]
36. Pulido N, Guevara-Morales JM, Rodriguez-López A, et al. 1H-Nuclear Magnetic Resonance Analysis of Urine as Diagnostic Tool for Organic Acidemias and Aminoacidopathies. Metabolites 2021;11:891. [Crossref] [PubMed]
37. Mendez KM, Reinke SN, Broadhurst DI. A comparative evaluation of the generalised predictive ability of eight machine learning algorithms across ten clinical metabolomics data sets for binary classification. Metabolomics 2019;15:150. [Crossref] [PubMed]
38. Sarvin B, Lagziel S, Sarvin N, et al. Fast and sensitive flow-injection mass spectrometry metabolomics by analyzing sample-specific ion distributions. Nat Commun 2020;11:3186. [Crossref] [PubMed]
39. Saudubray JM, Sedel F, Walter JH. Clinical approach to treatable inborn metabolic diseases: an introduction. J Inherit Metab Dis 2006;29:261-74. [Crossref] [PubMed]
40. Pazzi M, Colella S, Alladio E, et al. Statistical Optimization of Urinary Organic Acids Analysis by a Multi-Factorial Design of Experiment. Analytica 2020;1:14-23. [Crossref]
41. Roberts LD, Souza AL, Gerszten RE, et al. Targeted metabolomics. Curr Protoc Mol Biol 2012;Chapter 30:Unit 30.2.1-24.
42. Graham E, Lee J, Price M, et al. Integration of genomics and metabolomics for prioritization of rare disease variants: a 2018 literature review. J Inherit Metab Dis 2018;41:435-45. [Crossref] [PubMed]
43. Odom JD, Sutton VR. Metabolomics in Clinical Practice: Improving Diagnosis and Informing Management. Clin Chem 2021;67:1606-17. [Crossref] [PubMed]
44. Haijes HA, van der Ham M, Prinsen HCMT, et al. Untargeted Metabolomics for Metabolic Diagnostic Screening with Automated Data Interpretation Using a Knowledge-Based Algorithm. Int J Mol Sci 2020;21:979. [Crossref] [PubMed]
45. Thistlethwaite LR, Li X, Burrage LC, et al. Clinical diagnosis of metabolic disorders using untargeted metabolomic profiling and disease-specific networks learned from profiling data. Sci Rep 2022;12:6556. [Crossref] [PubMed]
46. de Sousa J, Vencálek O, Hron K, et al. Bayesian multiple hypotheses testing in compositional analysis of untargeted metabolomic data. Anal Chim Acta 2020;1097:49-61. [Crossref] [PubMed]
47. Ford L, Kennedy AD, Goodman KD, et al. Precision of a Clinical Metabolomics Profiling Platform for Use in the Identification of Inborn Errors of Metabolism. J Appl Lab Med 2020;5:342-56. [Crossref] [PubMed]
48. Kennedy AD, Pappan KL, Donti TR, et al. Elucidation of the complex metabolic profile of cerebrospinal fluid using an untargeted biochemical profiling assay. Mol Genet Metab 2017;121:83-90. [Crossref] [PubMed]
49. Shayota BJ, Donti TR, Xiao J, et al. Untargeted metabolomics as an unbiased approach to the diagnosis of inborn errors of metabolism of the non-oxidative branch of the pentose phosphate pathway. Mol Genet Metab 2020;131:147-54. [Crossref] [PubMed]
50. Donti TR, Cappuccio G, Hubert L, et al. Diagnosis of adenylosuccinate lyase deficiency by metabolomic profiling in plasma reveals a phenotypic spectrum. Mol Genet Metab Rep 2016;8:61-6. [Crossref] [PubMed]
51. Haijes HA, de Sain-van der Velden MGM, Prinsen HCMT, et al. Aspartylglycosamine is a biomarker for NGLY1-CDDG, a congenital disorder of deglycosylation. Mol Genet Metab 2019;127:368-72. [Crossref] [PubMed]
52. Liu H, Zhu J, Li Q, et al. Untargeted metabolomic analysis of urine samples for diagnosis of inherited metabolic disorders. Funct Integr Genomics 2021;21:645-53. [Crossref] [PubMed]
53. Sindelar M, Dyke JP, Deeb RS, et al. Untargeted Metabolite Profiling of Cerebrospinal Fluid Uncovers Biomarkers for Severity of Late Infantile Neuronal Ceroid Lipofuscinosis (CLN2, Batten Disease). Sci Rep 2018;8:15229. [Crossref] [PubMed]
54. Haijes HA, Willemsen M, Van der Ham M, et al. Direct Infusion Based Metabolomics Identifies Metabolic Disease in Patients' Dried Blood Spots and Plasma. Metabolites 2019;9:12. [Crossref] [PubMed]
55. Kerkhofs MHPM, Haijes HA, Willemsen AM, et al. Cross-Omics: Integrating Genomics with Metabolomics in Clinical Diagnostics. Metabolites 2020;10:206. [Crossref] [PubMed]
56. Anzmann AF, Pinto S, Busa V, et al. Multi-omics studies in cellular models of methylmalonic acidemia and propionic acidemia reveal dysregulation of serine metabolism. Biochim Biophys Acta Mol Basis Dis 2019;1865:165538. [Crossref] [PubMed]
57. Ruiz M, Jové M, Schlüter A, et al. Altered glycolipid and glycerophospholipid signaling drive inflammatory cascades in adrenomyeloneuropathy. Hum Mol Genet 2015;24:6861-76. [Crossref] [PubMed]
58. Sharma R, Reinstadler B, Engelstad K, et al. Circulating markers of NADH-reductive stress correlate with mitochondrial disease severity. J Clin Invest 2021;131:136055. [Crossref] [PubMed]
59. McDermott JE, Wang J, Mitchell H, et al. Challenges in Biomarker Discovery: Combining Expert Insights with Statistical Analysis of Complex Omics Data. Expert Opin Med Diagn 2013;7:37-51. [Crossref] [PubMed]
60. Schadewaldt P, Bodner-Leidecker A, Hammen HW, et al. Significance of L-alloisoleucine in plasma for diagnosis of maple syrup urine disease. Clin Chem 1999;45:1734-40. [Crossref] [PubMed]
61. Chakrapani A, Cleary MA, Wraith JE. Detection of inborn errors of metabolism in the newborn. Arch Dis Child Fetal Neonatal Ed 2001;84:F205-10. [Crossref] [PubMed]
62. Jain M, Kennedy AD, Elsea SH, et al. Analytes related to erythrocyte metabolism are reliable biomarkers for preanalytical error due to delayed plasma processing in metabolomics studies. Clin Chim Acta 2017;466:105-11. [Crossref] [PubMed]
63. Goodman K, Mitchell M, Evans AM, et al. Assessment of the effects of repeated freeze thawing and extended bench top processing of plasma samples using untargeted metabolomics. Metabolomics 2021;17:31. [Crossref] [PubMed]
64. Wangler MF, Hubert L, Donti TR, et al. A metabolomic map of Zellweger spectrum disorders reveals novel disease biomarkers. Genet Med 2018;20:1274-83. [Crossref] [PubMed]
65. Johnson CH, Ivanisevic J, Siuzdak G. Metabolomics: beyond biomarkers and towards mechanisms. Nat Rev Mol Cell Biol 2016;17:451-9. [Crossref] [PubMed]
66. Hoegen B, Hampstead JE, Engelke UFH, et al. Application of metabolite set enrichment analysis on untargeted metabolomics data prioritises relevant pathways and detects novel biomarkers for inherited metabolic disorders. J Inherit Metab Dis 2022;45:682-95. [Crossref] [PubMed]
67. Norman BP, Davison AS, Hughes JH, et al. Metabolomic studies in the inborn error of metabolism alkaptonuria reveal new biotransformations in tyrosine metabolism. Genes Dis 2022;9:1129-42. [Crossref] [PubMed]
68. Taylor Fischer S, Frederick AB, Tran V, et al. Metabolic perturbations in classic galactosemia beyond the Leloir pathway: Insights from an untargeted metabolomic study. J Inherit Metab Dis 2019;42:254-63. [Crossref] [PubMed]
69. Tebani A, Schmitz-Afonso I, Abily-Donval L, et al. Urinary metabolic phenotyping of mucopolysaccharidosis type I combining untargeted and targeted strategies with data modeling. Clin Chim Acta 2017;475:7-14. [Crossref] [PubMed]
70. Aref-Eshghi E, Bend EG, Hood RL, et al. BAFopathies' DNA methylation epi-signatures demonstrate diagnostic utility and functional continuum of Coffin-Siris and Nicolaides-Baraitser syndromes. Nat Commun 2018;9:4885. [Crossref] [PubMed]
71. Venter L, Lindeque Z, Jansen van Rensburg P, et al. Untargeted urine metabolomics reveals a biosignature for muscle respiratory chain deficiencies. Metabolomics 2015;11:111-21. [Crossref]
72. Mathis T, Poms M, Köfeler H, et al. Untargeted plasma metabolomics identifies broad metabolic perturbations in glycogen storage disease type I. J Inherit Metab Dis 2022;45:235-47. [Crossref] [PubMed]
73. McCoin CS, Piccolo BD, Knotts TA, et al. Unique plasma metabolomic signatures of individuals with inherited disorders of long-chain fatty acid oxidation. J Inherit Metab Dis 2016;39:399-408. [Crossref] [PubMed]
74. Sarode GV, Kim K, Kieffer DA, et al. Metabolomics profiles of patients with Wilson disease reveal a distinct metabolic signature. Metabolomics 2019;15:43. [Crossref] [PubMed]
75. Van Dooijeweert B, Broeks MH, Verhoeven-Duif NM, et al. Untargeted metabolic profiling in dried blood spots identifies disease fingerprint for pyruvate kinase deficiency. Haematologica 2021;106:2720-5. [Crossref] [PubMed]
76. 100,000 Genomes Project Pilot Investigators; Smedley D, Smith KR, et al. 100,000 Genomes Pilot on Rare-Disease Diagnosis in Health Care - Preliminary Report. N Engl J Med 2021;385:1868-80.
77. Pillai NR, Amin H, Gijavanekar C, et al. Hematologic presentation and the role of untargeted metabolomics analysis in monitoring treatment for riboflavin transporter deficiency. Am J Med Genet A 2020;182:2781-7. [Crossref] [PubMed]
78. Schoen MS, Singh RH. Plasma metabolomic profile changes in females with phenylketonuria following a camp intervention. Am J Clin Nutr 2022;115:811-21. [Crossref] [PubMed]
79. Lippi G, Plebani M. Integrated diagnostics: the future of laboratory medicine? Biochem Med (Zagreb) 2020;30:010501. [Crossref] [PubMed]
80. Seydel C. Single-cell metabolomics hits its stride. Nat Methods 2021;18:1452-6. [Crossref] [PubMed]
81. Petrick L, Edmands W, Schiffman C, et al. An untargeted metabolomics method for archived newborn dried blood spots in epidemiologic studies. Metabolomics 2017;13:27. [Crossref] [PubMed]
82. Knottnerus SJG, Pras-Raves ML, van der Ham M, et al. Prediction of VLCAD deficiency phenotype by a metabolic fingerprint in newborn screening bloodspots. Biochim Biophys Acta Mol Basis Dis 2020;1866:165725. [Crossref] [PubMed]
83. Skogvold HB, Sandås EM, Østeby A, et al. Bridging the Polar and Hydrophobic Metabolome in Single-Run Untargeted Liquid Chromatography-Mass Spectrometry Dried Blood Spot Metabolomics for Clinical Purposes. J Proteome Res 2021;20:4010-21. [Crossref] [PubMed]
84. Watson MS. Current status of newborn screening: decision-making about the conditions to include in screening programs. Ment Retard Dev Disabil Res Rev 2006;12:230-5. [Crossref] [PubMed]
85. Woerner AC, Gallagher RC, Vockley J, et al. The Use of Whole Genome and Exome Sequencing for Newborn Screening: Challenges and Opportunities for Population Health. Front Pediatr 2021;9:663752. [Crossref] [PubMed]
86. Roman TS, Crowley SB, Roche MI, et al. Genomic Sequencing for Newborn Screening: Results of the NC NEXUS Project. Am J Hum Genet 2020;107:596-611. [Crossref] [PubMed]
87. Pinto FG, Mahmud I, Harmon TA, et al. Rapid Prostate Cancer Noninvasive Biomarker Screening Using Segmented Flow Mass Spectrometry-Based Untargeted Metabolomics. J Proteome Res 2020;19:2080-91. [Crossref] [PubMed]
88. van Karnebeek CD, Bonafé L, Wen XY, et al. NANS-mediated synthesis of sialic acid is required for brain and skeletal development. Nat Genet 2016;48:777-84. [Crossref] [PubMed]
89. Alaimo JT, Glinton KE, Liu N, et al. Integrated analysis of metabolomic profiling and exome data supplements sequence variant interpretation, classification, and diagnosis. Genet Med 2020;22:1560-6. [Crossref] [PubMed]
90. Bongaerts M, Bonte R, Demirdas S, et al. Integration of metabolomics with genomics: Metabolic gene prioritization using metabolomics data and genomic variant (CADD) scores. Mol Genet Metab 2022;136:199-218. [Crossref] [PubMed]
91. Graham Linck EJ, Richmond PA, Tarailo-Graovac M, et al. metPropagate: network-guided propagation of metabolomic information for prioritization of metabolic disease genes. NPJ Genom Med 2020;5:25. [Crossref] [PubMed]
92. Kennedy AD, Pappan KL, Donti T, et al. 2-Pyrrolidinone and Succinimide as Clinical Screening Biomarkers for GABA-Transaminase Deficiency: Anti-seizure Medications Impact Accurate Diagnosis. Front Neurosci 2019;13:394. [Crossref] [PubMed]
93. Brnich SE, Abou Tayoun AN, Couch FJ, et al. Recommendations for application of the functional evidence PS3/BS3 criterion using the ACMG/AMP sequence variant interpretation framework. Genome Med 2019;12:3. [Crossref] [PubMed]
94. Richards S, Aziz N, Bale S, et al. Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology. Genet Med 2015;17:405-24. [Crossref] [PubMed]
95. Koelmel JP, Ulmer CZ, Jones CM, et al. Common cases of improper lipid annotation using high-resolution tandem mass spectrometry data and corresponding limitations in biological interpretation. Biochim Biophys Acta Mol Cell Biol Lipids 2017;1862:766-70. [Crossref] [PubMed]
96. Dhiman N, Hall L, Wohlfiel SL, et al. Performance and cost analysis of matrix-assisted laser desorption ionization-time of flight mass spectrometry for routine identification of yeast. J Clin Microbiol 2011;49:1614-6. [Crossref] [PubMed]
97. Zhang A, Sun H, Xu H, et al. Cell metabolomics. OMICS 2013;17:495-501. [Crossref] [PubMed]
98. Courraud J, Ernst M, Svane Laursen S, et al. Studying Autism Using Untargeted Metabolomics in Newborn Screening Samples. J Mol Neurosci 2021;71:1378-93. [Crossref] [PubMed]
99. Peters TMA, Engelke UFH, de Boer S, et al. Confirmation of neurometabolic diagnoses using age-dependent cerebrospinal fluid metabolomic profiles. J Inherit Metab Dis 2020;43:1112-20. [Crossref] [PubMed]
100. Bongaerts M, Bonte R, Demirdas S, et al. Using Out-of-Batch Reference Populations to Improve Untargeted Metabolomics for Screening Inborn Errors of Metabolism. Metabolites 2020;11:8. [Crossref] [PubMed]
101. Biagini D, Lomonaco T, Ghimenti S, et al. Using labelled internal standards to improve needle trap micro-extraction technique prior to gas chromatography/mass spectrometry. Talanta 2019;200:145-55. [Crossref] [PubMed]
102. Yang J, Zhao X, Lu X, et al. A data preprocessing strategy for metabolomics to reduce the mask effect in data analysis. Front Mol Biosci 2015;2:4. [Crossref] [PubMed]
103. Beebe K, Kennedy AD. Sharpening Precision Medicine by a Thorough Interrogation of Metabolic Individuality. Comput Struct Biotechnol J 2016;14:97-105. [Crossref] [PubMed]
104. Popejoy AB, Ritter DI, Crooks K, et al. The clinical imperative for inclusivity: Race, ethnicity, and ancestry (REA) in genomics. Hum Mutat 2018;39:1713-20. [Crossref] [PubMed]
105. Landry LG, Ali N, Williams DR, et al. Lack Of Diversity In Genomic Databases Is A Barrier To Translating Precision Medicine Research Into Practice. Health Aff (Millwood) 2018;37:780-5. [Crossref] [PubMed]
106. Rosenfeld JA, Mason CE, Smith TM. Limitations of the human reference genome for personalized genomics. PLoS One 2012;7:e40294. [Crossref] [PubMed]
107. Hudson JF, Phelan MM, Owens DJ, et al. "Fuel for the Damage Induced": Untargeted Metabolomics in Elite Rugby Union Match Play. Metabolites 2021;11:544. [Crossref] [PubMed]
108. Wishart DS, Guo A, Oler E, et al. HMDB 5.0: the Human Metabolome Database for 2022. Nucleic Acids Res 2022;50:D622-31. [Crossref] [PubMed]
109. Leung DG, Cohen JS, Michelle EH, et al. Mitochondrial DNA Deletions With Low-Level Heteroplasmy in Adult-Onset Myopathy. J Clin Neuromuscul Dis 2018;19:117-23. [Crossref] [PubMed]
110. Sharer JD, Bodamer O, Longo N, et al. Laboratory diagnosis of creatine deficiency syndromes: a technical standard and guideline of the American College of Medical Genetics and Genomics. Genet Med 2017;19:256-63. [Crossref] [PubMed]
111. Almontashiri NAM, Zha L, Young K, et al. Clinical Validation of Targeted and Untargeted Metabolomics Testing for Genetic Disorders: A 3 Year Comparative Study. Sci Rep 2020;10:9382. [Crossref] [PubMed]
112. Barbitoff YA, Polev DE, Glotov AS, et al. Systematic dissection of biases in whole-exome and whole-genome sequencing reveals major determinants of coding sequence coverage. Sci Rep 2020;10:2057. [Crossref] [PubMed]
113. Steinbusch LKM, Wang P, Waterval HWAH, et al. Targeted urine metabolomics with a graphical reporting tool for rapid diagnosis of inborn errors of metabolism. J Inherit Metab Dis 2021;44:1113-23. [Crossref] [PubMed]