Application of a Computational Metabolomics Workflow for the Diagnosis of Inborn Errors of Metabolism in a Laboratory Setting.

Inborn errors of metabolism constitute a set of hereditary diseases that impose severe medical and physical challenges in the affected individual, in particular, for the pediatric patient population. Timely diagnosis is crucial for these patients, as any delay could result in irreversible health damage, underscoring the importance of early initiation of personalized treatment. Current routine diagnostic screening for inborn errors of metabolism relies on various targeted analyses of established biomarkers. However, this approach is time-consuming, focuses on a limited number of tests (based on clinical information) with a relatively small number of biomarkers, and does not facilitate the identification of new markers. In contrast, untargeted metabolomics-based screening offers a more efficient diagnostic solution, by assessing thousands of metabolites across multiple metabolic pathways in a single test. This not only saves time but also conserves resources for clinicians, the diagnostic laboratory, and for patients.This chapter describes the computational workflow of our "Next Generation Metabolic Screening" approach, which is a metabolomics-based method that is currently applied at the Translational Metabolic Laboratory of the Radboud University Medical Center (the Netherlands) for the diagnosis of inborn errors of metabolism.

A Delphi Survey Study to Formulate Statements on the Treatability of Inherited Metabolic Disorders to Decide on Eligibility for Newborn Screening.

The Wilson and Jungner (W&J) and Andermann criteria are meant to help select diseases eligible for population-based screening. With the introduction of next-generation sequencing (NGS) methods for newborn screening (NBS), more inherited metabolic diseases (IMDs) can technically be included, and a revision of the criteria was attempted. This study aimed to formulate statements and investigate whether those statements could elaborate on the criterion of treatability for IMDs to decide on eligibility for NBS. An online Delphi study was started among a panel of Dutch IMD experts (EPs). EPs evaluated, amended, and approved statements on treatability that were subsequently applied to 10 IMDs. After two rounds of Delphi, consensus was reached on 10 statements. Application of these statements selected 5 out of 10 IMDs proposed for this study as eligible for NBS, including 3 IMDs in the current Dutch NBS. The statement: 'The expected benefit/burden ratio of early treatment is positive and results in a significant health outcome' contributed most to decision-making. Our Delphi study resulted in 10 statements that can help to decide on eligibility for inclusion in NBS based on treatability, also showing that other criteria could be handled in a comparable way. Validation of the statements is required before these can be applied as guidance to authorities.

Benchmarking Outlier Detection Methods for Detecting IEM Patients in Untargeted Metabolomics Data.

Untargeted metabolomics (UM) is increasingly being deployed as a strategy for screening patients that are suspected of having an inborn error of metabolism (IEM). In this study, we examined the potential of existing outlier detection methods to detect IEM patient profiles. We benchmarked 30 different outlier detection methods when applied to three untargeted metabolomics datasets. Our results show great differences in IEM detection performances across the various methods. The methods DeepSVDD and R-graph performed most consistently across the three metabolomics datasets. For datasets with a more balanced number of samples-to-features ratio, we found that AE reconstruction error, Mahalanobis and PCA reconstruction error also performed well. Furthermore, we demonstrated the importance of a PCA transform prior to applying an outlier detection method since we observed that this increases the performance of several outlier detection methods. For only one of the three metabolomics datasets, we observed clinically satisfying performances for some outlier detection methods, where we were able to detect 90% of the IEM patient samples while detecting no false positives. These results suggest that outlier detection methods have the potential to aid the clinical investigator in routine screening for IEM using untargeted metabolomics data, but also show that further improvements are needed to ensure clinically satisfying performances.

Diagnosing, discarding, or de-VUSsing: A practical guide to (un)targeted metabolomics as variant-transcending functional tests.

PURPOSE: For patients with inherited metabolic disorders (IMDs), any diagnostic delay should be avoided because early initiation of personalized treatment could prevent irreversible health damage. To improve diagnostic interpretation of genetic data, gene function tests can be valuable assets. For IMDs, variant-transcending functional tests are readily available through (un)targeted metabolomics assays. To support the application of metabolomics for this purpose, we developed a gene-based guide to select functional tests to either confirm or exclude an IMD diagnosis. METHODS: Using information from a diagnostic IMD exome panel, Kyoto Encyclopedia of Genes and Genomes, and Inborn Errors of Metabolism Knowledgebase, we compiled a guide for metabolomics-based gene function tests. From our practical experience with this guide, we retrospectively selected illustrative cases for whom combined metabolomic/genomic testing improved diagnostic success and evaluated the effect hereof on clinical management. RESULTS: The guide contains 2047 metabolism-associated genes for which a validated or putative variant-transcending gene function test is available. We present 16 patients for whom metabolomic testing either confirmed or ruled out the presence of a second pathogenic variant, validated or ruled out pathogenicity of variants of uncertain significance, or identified a diagnosis initially missed by genetic analysis. CONCLUSION: Metabolomics-based gene function tests provide additional value in the diagnostic trajectory of patients with suspected IMD by enhancing and accelerating diagnostic success.

Long-term MRI Findings in Patients With Cerebrotendinous Xanthomatosis Treated With Chenodeoxycholic Acid.

OBJECTIVES: To describe long-term follow-up brain MRI findings in patients with cerebrotendinous xanthomatosis (CTX) treated with chenodeoxycholic acid (CDCA). METHODS: Of a cohort of 79 Dutch patients with CTX, we retrospectively reviewed brain MRI findings of patients at diagnosis (before the start of treatment) and after long-term follow-up (7-27 years) in 12 patients. In addition, we report on 2 families with remarkable brain MRI findings. RESULTS: MRI abnormalities showed progression in all 7 patients diagnosed at 24 years or older and only in 1 of 5 patients diagnosed younger than 24 years. MRI findings in the other patients diagnosed younger than 24 years were normal at baseline and remained normal even after follow-up of more than 25 years. The total MRI scores at baseline were 2 and 19 and at follow-up 4 and 37, respectively, for patients diagnosed before or after the age of 24 years, despite a comparable number of treatment years. DISCUSSION: MRI findings are fully in line with our long-term treatment effect article, emphasizing the importance of early diagnosis and treatment in CTX. Expanding the spectrum of brain MRI findings (including the finding of a posterior leukoencephalopathy) leads to a better understanding of the heterogeneity of this treatable disease.

Neonatal Long-Chain 3-Ketoacyl-CoA Thiolase deficiency: Clinical-biochemical phenotype, sodium-D,L-3-hydroxybutyrate treatment experience and cardiac evaluation using speckle echocardiography.

Isolated long-chain 3-keto-acyl CoA thiolase (LCKAT) deficiency is a rare long-chain fatty acid oxidation disorder caused by mutations in HADHB. LCKAT is part of a multi-enzyme complex called the mitochondrial trifunctional protein (MTP) which catalyzes the last three steps in the long-chain fatty acid oxidation. Until now, only three cases of isolated LCKAT deficiency have been described. All patients developed a severe cardiomyopathy and died before the age of 7 weeks. Here, we describe a newborn with isolated LCKAT deficiency, presenting with neonatal-onset cardiomyopathy, rhabdomyolysis, hypoglycemia and lactic acidosis. Bi-allelic 185G > A (p.Arg62His) and c1292T > C (p.Phe431Ser) mutations were found in HADHB. Enzymatic analysis in both lymphocytes and cultured fibroblasts revealed LCKAT deficiency with a normal long-chain 3-hydroxyacyl-CoA dehydrogenase (LCHAD, also part of MTP) enzyme activity. Clinically, the patient showed recurrent cardiomyopathy, which was monitored by speckle tracking echocardiography. Subsequent treatment with special low-fat formula, low in long chain triglycerides (LCT) and supplemented with medium chain triglycerides (MCT) and ketone body therapy in (sodium-D,L-3-hydroxybutyrate) was well tolerated and resulted in improved carnitine profiles and cardiac function. Resveratrol, a natural polyphenol that has been shown to increase fatty acid oxidation, was also considered as a potential treatment option but showed no in vitro benefits in the patient's fibroblasts. Even though our patient deceased at the age of 13 months, early diagnosis and prompt initiation of dietary management with addition of sodium-D,L-3-hydroxybutyrate may have contributed to improved cardiac function and a much longer survival when compared to the previously reported cases of isolated LCKAT-deficiency.

Preanalytical Pitfalls in Untargeted Plasma Nuclear Magnetic Resonance Metabolomics of Endocrine Hypertension.

Despite considerable morbidity and mortality, numerous cases of endocrine hypertension (EHT) forms, including primary aldosteronism (PA), pheochromocytoma and functional paraganglioma (PPGL), and Cushing's syndrome (CS), remain undetected. We aimed to establish signatures for the different forms of EHT, investigate potentially confounding effects and establish unbiased disease biomarkers. Plasma samples were obtained from 13 biobanks across seven countries and analyzed using untargeted NMR metabolomics. We compared unstratified samples of 106 PHT patients to 231 EHT patients, including 104 PA, 94 PPGL and 33 CS patients. Spectra were subjected to a multivariate statistical comparison of PHT to EHT forms and the associated signatures were obtained. Three approaches were applied to investigate and correct confounding effects. Though we found signatures that could separate PHT from EHT forms, there were also key similarities with the signatures of sample center of origin and sample age. The study design restricted the applicability of the corrections employed. With the samples that were available, no biomarkers for PHT vs. EHT could be identified. The complexity of the confounding effects, evidenced by their robustness to correction approaches, highlighted the need for a consensus on how to deal with variabilities probably attributed to preanalytical factors in retrospective, multicenter metabolomics studies.

Pre- versus post-operative untargeted plasma nuclear magnetic resonance spectroscopy metabolomics of pheochromocytoma and paraganglioma.

PURPOSE: Pheochromocytomas and Paragangliomas (PPGL) result in chronic catecholamine excess and serious health complications. A recent study obtained a metabolic signature in plasma from PPGL patients; however, its targeted nature may have generated an incomplete picture and a broader approach could provide additional insights. We aimed to characterize the plasma metabolome of PPGL patients before and after surgery, using an untargeted approach, and to broaden the scope of the investigated metabolic impact of these tumors. DESIGN: A cohort of 36 PPGL patients was investigated. Blood plasma samples were collected before and after surgical tumor removal, in association with clinical and tumor characteristics. METHODS: Plasma samples were analyzed using untargeted nuclear magnetic resonance (NMR) spectroscopy metabolomics. The data were evaluated using a combination of uni- and multi-variate statistical methods. RESULTS: Before surgery, patients with a nonadrenergic tumor could be distinguished from those with an adrenergic tumor based on their metabolic profiles. Tyrosine levels were significantly higher in patients with high compared to those with low BMI. Comparing subgroups of pre-operative samples with their post-operative counterparts, we found a metabolic signature that included ketone bodies, glucose, organic acids, methanol, dimethyl sulfone and amino acids. Three signals with unclear identities were found to be affected. CONCLUSIONS: Our study suggests that the pathways of glucose and ketone body homeostasis are affected in PPGL patients. BMI-related metabolite levels were also found to be altered, potentially linking muscle atrophy to PPGL. At baseline, patient metabolomes could be discriminated based on their catecholamine phenotype.

Metabolite Identification Using Infrared Ion Spectroscopy─Novel Biomarkers for Pyridoxine-Dependent Epilepsy.

Untargeted liquid chromatography-mass spectrometry (LC-MS)-based metabolomics strategies are being increasingly applied in metabolite screening for a wide variety of medical conditions. The long-standing "grand challenge" in the utilization of this approach is metabolite identification─confidently determining the chemical structures of m/z-detected unknowns. Here, we use a novel workflow based on the detection of molecular features of interest by high-throughput untargeted LC-MS analysis of patient body fluids combined with targeted molecular identification of those features using infrared ion spectroscopy (IRIS), effectively providing diagnostic IR fingerprints for mass-isolated targets. A significant advantage of this approach is that in silico-predicted IR spectra of candidate chemical structures can be used to suggest the molecular structure of unknown features, thus mitigating the need for the synthesis of a broad range of physical reference standards. Pyridoxine-dependent epilepsy (PDE-ALDH7A1) is an inborn error of lysine metabolism, resulting from a mutation in the ALDH7A1 gene that leads to an accumulation of toxic levels of α-aminoadipic semialdehyde (α-AASA), piperideine-6-carboxylate (P6C), and pipecolic acid in body fluids. While α-AASA and P6C are known biomarkers for PDE in urine, their instability makes them poor candidates for diagnostic analysis from blood, which would be required for application in newborn screening protocols. Here, we use combined untargeted metabolomics-IRIS to identify several new biomarkers for PDE-ALDH7A1 that can be used for diagnostic analysis in urine, plasma, and cerebrospinal fluids and that are compatible with analysis in dried blood spots for newborn screening. The identification of these novel metabolites has directly provided novel insights into the pathophysiology of PDE-ALDH7A1.

Metabolomics-Based Screening of Inborn Errors of Metabolism: Enhancing Clinical Application with a Robust Computational Pipeline.

Inborn errors of metabolism (IEM) are inherited conditions caused by genetic defects in enzymes or cofactors. These defects result in a specific metabolic fingerprint in patient body fluids, showing accumulation of substrate or lack of an end-product of the defective enzymatic step. Untargeted metabolomics has evolved as a high throughput methodology offering a comprehensive readout of this metabolic fingerprint. This makes it a promising tool for diagnostic screening of IEM patients. However, the size and complexity of metabolomics data have posed a challenge in translating this avalanche of information into knowledge, particularly for clinical application. We have previously established next-generation metabolic screening (NGMS) as a metabolomics-based diagnostic tool for analyzing plasma of individual IEM-suspected patients. To fully exploit the clinical potential of NGMS, we present a computational pipeline to streamline the analysis of untargeted metabolomics data. This pipeline allows for time-efficient and reproducible data analysis, compatible with ISO:15189 accredited clinical diagnostics. The pipeline implements a combination of tools embedded in a workflow environment for large-scale clinical metabolomics data analysis. The accompanying graphical user interface aids end-users from a diagnostic laboratory for efficient data interpretation and reporting. We also demonstrate the application of this pipeline with a case study and discuss future prospects.

Abnormal VLCADD newborn screening resembling MADD in four neonates with decreased riboflavin levels and VLCAD activity.

Early detection of congenital disorders by newborn screening (NBS) programs is essential to prevent or limit disease manifestation in affected neonates. These programs balance between the detection of the highest number of true cases and the lowest number of false-positives. In this case report, we describe four unrelated cases with a false-positive NBS result for very-long-chain acyl-CoA dehydrogenase deficiency (VLCADD). Three neonates presented with decreased but not deficient VLCAD enzyme activity and two of them carried a single heterozygous ACADVL c.1844G>A mutation. Initial biochemical investigations after positive NBS referral in these infants revealed acylcarnitine and organic acid profiles resembling those seen in multiple acyl-CoA dehydrogenase deficiency (MADD). Genetic analysis did not reveal any pathogenic mutations in the genes encoding the electron transfer flavoprotein (ETF alpha and beta subunits) nor in ETF dehydrogenase. Subsequent further diagnostics revealed decreased levels of riboflavin in the newborns and oral riboflavin administration normalized the MADD-like biochemical profiles. During pregnancy, the mothers followed a vegan, vegetarian or lactose-free diet which probably caused alimentary riboflavin deficiency in the neonates. This report demonstrates that a secondary (alimentary) maternal riboflavin deficiency in combination with reduced VLCAD activity in the newborns can result in an abnormal VLCADD/MADD acylcarnitine profile and can cause false-positive NBS. We hypothesize that maternal riboflavin deficiency contributed to the false-positive VLCADD neonatal screening results.

NANS-CDG: Delineation of the Genetic, Biochemical, and Clinical Spectrum.

Background: NANS-CDG is a recently described congenital disorder of glycosylation caused by biallelic genetic variants in NANS, encoding an essential enzyme in de novo sialic acid synthesis. Sialic acid at the end of glycoconjugates plays a key role in biological processes such as brain and skeletal development. Here, we present an observational cohort study to delineate the genetic, biochemical, and clinical phenotype and assess possible correlations. Methods: Medical and laboratory records were reviewed with retrospective extraction and analysis of genetic, biochemical, and clinical data (2016-2020). Results: Nine NANS-CDG patients (nine families, six countries) referred to the Radboudumc CDG Center of Expertise were included. Phenotyping confirmed the hallmark features including intellectual developmental disorder (IDD) (n = 9/9; 100%), facial dysmorphisms (n = 9/9; 100%), neurologic impairment (n = 9/9; 100%), short stature (n = 8/9; 89%), skeletal dysplasia (n = 8/9; 89%), and short limbs (n = 8/9; 89%). Newly identified features include ophthalmological abnormalities (n = 6/9; 67%), an abnormal septum pellucidum (n = 6/9; 67%), (progressive) cerebral atrophy and ventricular dilatation (n = 5/9; 56%), gastrointestinal dysfunction (n = 5/9; 56%), thrombocytopenia (n = 5/9; 56%), and hypo-low-density lipoprotein cholesterol (n = 4/9; 44%). Biochemically, elevated urinary excretion of N-acetylmannosamine (ManNAc) is pathognomonic, the concentrations of which show a significant correlation with clinical severity. Genotypically, eight novel NANS variants were identified. Three severely affected patients harbored identical compound heterozygous pathogenic variants, one of whom was initiated on experimental prenatal and postnatal treatment with oral sialic acid. This patient showed markedly better psychomotor development than the other two genotypically identical males. Conclusions: ManNAc screening should be considered in all patients with IDD, short stature with short limbs, facial dysmorphisms, neurologic impairment, and an abnormal septum pellucidum +/- congenital and neurodegenerative lesions on brain imaging, to establish a precise diagnosis and contribute to prognostication. Personalized management includes accurate genetic counseling and access to proper supports and tailored care for gastrointestinal symptoms, thrombocytopenia, and epilepsy, as well as rehabilitation services for cognitive and physical impairments. Motivated by the short-term positive effects of experimental treatment with oral sialic, we have initiated this intervention with protocolized follow-up of neurologic, systemic, and growth outcomes in four patients. Research is ongoing to unravel pathophysiology and identify novel therapeutic targets.

Monitoring phenylalanine concentrations in the follow-up of phenylketonuria patients: An inventory of pre-analytical and analytical variation.

BACKGROUND: Reliable measurement of phenylalanine (Phe) is a prerequisite for adequate follow-up of phenylketonuria (PKU) patients. However, previous studies have raised concerns on the intercomparability of plasma and dried blood spot (DBS) Phe results. In this study, we made an inventory of differences in (pre-)analytical methodology used for Phe determination across Dutch laboratories, and compared DBS and plasma results. METHODS: Through an online questionnaire, we assessed (pre-)analytical Phe measurement procedures of seven Dutch metabolic laboratories. To investigate the difference between plasma and DBS Phe, participating laboratories received simultaneously collected plasma-DBS sets from 23 PKU patients. In parallel, 40 sample sets of DBS spotted from either venous blood or capillary fingerprick were analyzed. RESULTS: Our data show that there is no consistency on standard operating procedures for Phe measurement. The association of DBS to plasma Phe concentration exhibits substantial inter-laboratory variation, ranging from a mean difference of -15.5% to +30.6% between plasma and DBS Phe concentrations. In addition, we found a mean difference of +5.8% in Phe concentration between capillary DBS and DBS prepared from venous blood. CONCLUSIONS: The results of our study point to substantial (pre-)analytical variation in Phe measurements, implicating that bloodspot Phe results should be interpreted with caution, especially when no correction factor is applied. To minimize variation, we advocate pre-analytical standardization and analytical harmonization of Phe measurements, including consensus on application of a correction factor to adjust DBS Phe to plasma concentrations.

Amadori rearrangement products as potential biomarkers for inborn errors of amino-acid metabolism.

The identification of disease biomarkers plays a crucial role in developing diagnostic strategies for inborn errors of metabolism and understanding their pathophysiology. A primary metabolite that accumulates in the inborn error phenylketonuria is phenylalanine, however its levels do not always directly correlate with clinical outcomes. Here we combine infrared ion spectroscopy and NMR spectroscopy to identify the Phe-glucose Amadori rearrangement product as a biomarker for phenylketonuria. Additionally, we find analogous amino acid-glucose metabolites formed in the body fluids of patients accumulating methionine, lysine, proline and citrulline. Amadori rearrangement products are well-known intermediates in the formation of advanced glycation end-products and have been associated with the pathophysiology of diabetes mellitus and ageing, but are now shown to also form under conditions of aminoacidemia. They represent a general class of metabolites for inborn errors of amino acid metabolism that show potential as biomarkers and may provide further insight in disease pathophysiology.

The role of clinical response to treatment in determining pathogenicity of genomic variants.

PURPOSE: The 2015 American College of Medical Genetics and Genomics/Association for Molecular Pathology (ACMG/AMP) guidelines for the interpretation of sequence variants provide a framework to standardize terminology in the classification of variants uncovered through genetic testing. We aimed to assess the validity of utilizing clinical response to therapies specifically targeted to a suspected disease in clarifying variant pathogenicity. METHODS: Five families with disparate clinical presentations and different genetic diseases evaluated and treated in multiple diagnostic settings are summarized. RESULTS: Extended evaluations indicated possible genetic diagnoses and assigned candidate causal variants, but the cumulative clinical, biochemical, and molecular information in each instance was not completely consistent with the identified disease. Initiation of treatment specific to the suspected diagnoses in the affected individuals led to clinical improvement in all five families. CONCLUSION: We propose that the effect of therapies that are specific and targeted to treatable genetic diseases embodies an in vivo physiological response and could be considered as additional criteria within the 2015 ACMG/AMP guidelines in determining genomic variant pathogenicity.

A newborn screening approach to diagnose 3-hydroxy-3-methylglutaryl-CoA lyase deficiency.

3-Hydroxy-3-methylglutaryl-coenzyme A lyase deficiency (HMGCLD) is a rare autosomal recessively inherited metabolic disorder. Patients suffer from avoidable neurologically devastating metabolic decompensations and thus would benefit from newborn screening (NBS). The diagnosis is currently made by measuring dry blood spot acylcarnitines (C5OH and C6DC) followed by urinary organic acid profiling for the differential diagnosis from several other disorders. Using untargeted metabolomics (reversed-phase UHPLC coupled to an Orbitrap Elite hybrid mass spectrometer) of plasma samples from 5 HMGCLD patients and 19 age-matched controls, we found 3-methylglutaconic acid and 3-hydroxy-3-methylglutaric acid, together with 3-hydroxyisovalerylcarnitine as the most discriminating metabolites between the groups. In order to evaluate the NBS potential of these metabolites we quantified the most discriminating metabolites from untargeted metabolomics in 23 blood spots from 4 HMGCLD patients and 55 controls by UHPLC tandem mass spectrometry. The results provide a tool for expanded NBS of HMGCLD using tandem mass spectrometry. Selected reaction monitoring transition 262/85 could be used in a first-tier NBS analysis to screen for elevated 3-hydroxyisovalerylcarnitine. In a positive case, a second-tier analysis of 3-hydroxy-3-methylglutaric acid and 3-methylglutaconic acid in a dry blood spot using UHPLC tandem mass spectrometry instruments confirms the diagnosis. In conclusion, we describe the identification of new diagnostic biomarkers for HMGCLD and their application in NBS in dry blood spots. By using second-tier testing, all patients with HMGCLD were unequivocally and correctly diagnosed.

Evaluation of 11 years of newborn screening for maple syrup urine disease in the Netherlands and a systematic review of the literature: Strategies for optimization.

Maple syrup urine disease (MSUD) leads to severe neurological deterioration unless diagnosed early and treated immediately. We have evaluated the effectiveness of 11 years of MSUD newborn screening (NBS) in the Netherlands (screening >72 hours, referral if both total leucine (Xle) and valine ≥400 µmol/L blood) and have explored possibilities for improvement by combining our data with a systematic literature review and data from Collaborative Laboratory Integrated Reports (CLIR). Dutch MSUD NBS characteristics and accuracy were determined. The hypothetical referral numbers in the Dutch population of additional screening markers suggested by CLIR were calculated. In a systematic review, articles reporting NBS leucine concentrations of confirmed patients were included. Our data showed that NBS of 1 963 465 newborns identified 4 MSUD patients and led to 118 false-positive referrals (PPV 3.28%; incidence 1:491 000 newborns). In literature, leucine is the preferred NBS parameter. Total leucine (Xle) concentrations (mass-spectrometry) of 53 detected and 8 false-negative patients (sampling age within 25 hours in 3 patients) reported in literature ranged from 288 to 3376 (median 900) and 42 to 325 (median 209) µmol/L blood respectively. CLIR showed increasing Xle concentrations with sampling age and early NBS sampling and milder variant MSUD phenotypes with (nearly) normal biochemical profiles are causes of false-negative NBS results. We evaluated the effect of additional screening markers and established the Xle/phenylalanine ratio as a promising additional marker ratio for increasing the PPV, while maintaining high sensitivity in the Dutch MSUD NBS.

metPropagate: network-guided propagation of metabolomic information for prioritization of metabolic disease genes.

Many inborn errors of metabolism (IEMs) are amenable to treatment, therefore early diagnosis is imperative. Whole-exome sequencing (WES) variant prioritization coupled with phenotype-guided clinical and bioinformatics expertise is typically used to identify disease-causing variants; however, it can be challenging to identify the causal candidate gene when a large number of rare and potentially pathogenic variants are detected. Here, we present a network-based approach, metPropagate, that uses untargeted metabolomics (UM) data from a single patient and a group of controls to prioritize candidate genes in patients with suspected IEMs. We validate metPropagate on 107 patients with IEMs diagnosed in Miller et al. (2015) and 11 patients with both CNS and metabolic abnormalities. The metPropagate method ranks candidate genes by label propagation, a graph-smoothing algorithm that considers each gene's metabolic perturbation in addition to the network of interactions between neighbors. metPropagate was able to prioritize at least one causative gene in the top 20th percentile of candidate genes for 92% of patients with known IEMs. Applied to patients with suspected neurometabolic disease, metPropagate placed at least one causative gene in the top 20th percentile in 9/11 patients, and ranked the causative gene more highly than Exomiser's phenotype-based ranking in 6/11 patients. Interestingly, ranking by a weighted combination of metPropagate and Exomiser scores resulted in improved prioritization. The results of this study indicate that network-based analysis of UM data can provide an additional mode of evidence to prioritize causal genes in patients with suspected IEMs.