Tag: metabolomics

Metabolomics-Based Machine Learning Diagnostics of Post-Acute Sequelae of SARS-CoV-2 Infection

Abstract:

Background: COVID-19 has taken millions of lives and continues to affect people worldwide. Post-Acute Sequelae of SARS-CoV-2 Infection (also known as Post-Acute Sequelae of COVID-19 (PASC) or more commonly, Long COVID) occurs in the aftermath of COVID-19 and is poorly understood despite its widespread effects.

Methods: We created a machine-learning model that distinguishes PASC from PASC-similar diseases. The model was trained to recognize PASC-dysregulated metabolites (p ≤ 0.05) using molecular descriptors.

Results: Our multi-layer perceptron model accurately recognizes PASC-dysregulated metabolites in the independent testing set, with an AUC-ROC of 0.8991, and differentiates PASC from myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS), Lyme disease, postural orthostatic tachycardia syndrome (POTS), and irritable bowel syndrome (IBS). However, it was unable to differentiate fibromyalgia (FM) from PASC.

Conclusions: By creating and testing models pairwise on each of these diseases, we elucidated the unique strength of the similarity between FM and PASC relative to other PASC-similar diseases. Our approach is unique to PASC diagnosis, and our use of molecular descriptors enables our model to work with any metabolite where molecular descriptors can be identified, as these descriptors can be generated and compared for any metabolite. Our study presents a novel approach to PASC diagnosis that partially circumvents the lengthy process of exclusion, potentially facilitating faster interventions and improved patient outcomes.

Source: Cai E, Kouznetsova VL, Tsigelny IF. Metabolomics-Based Machine Learning Diagnostics of Post-Acute Sequelae of SARS-CoV-2 Infection. Metabolites. 2025 Dec 17;15(12):801. doi: 10.3390/metabo15120801. PMID: 41441042; PMCID: PMC12734907. https://pmc.ncbi.nlm.nih.gov/articles/PMC12734907/ (Full text)

Leveraging Explainable Automated Machine Learning (AutoML) and Metabolomics for Robust Diagnosis and Pathophysiological Insights in Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS)

Abstract:

Background/Objectives: Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS) is a debilitating complex disease with an elusive etiology, lacking objective diagnostic biomarkers. This study leverages advanced Automated Machine Learning (AutoML) to analyze plasma metabolomic and lipidomic profiles for the purpose of ME/CFS detection.

Methods: We utilized a publicly available dataset comprising 888 metabolic features from 106 ME/CFS patients and 91 matched controls. Three AutoML frameworks-TPOT, Auto-Sklearn, and H2O AutoML-were benchmarked under identical time constraints. Univariate ROC and PLS-DA analyses with cross-validation, permutation testing, and VIP-based feature selection were applied to standardized, log-transformed omics data to identify significant discriminatory metabolites/lipids and assess their intercorrelations.

Results: TPOT significantly outperformed its counterparts, achieving an area under the curve (AUC) of 92.1%, accuracy of 87.3%, sensitivity of 85.8%, and specificity of 89.0%. The PLS-DA model revealed a moderate but statistically significant discrimination between ME/CFS and controls. Explainable artificial intelligence (XAI) via SHAP analysis of the optimal TPOT model identified key metabolites implicating dysregulated pathways in mitochondrial energy metabolism (succinic acid, pyruvic acid, leucine), chronic inflammation (prostaglandin D2, 11,12-EET), gut-brain axis communication (glycocholic acid), and cell membrane integrity (pc(35:2)a).

Conclusions: Our results demonstrate that TPOT-derived models not only provide a highly accurate and robust diagnostic tool but also yield biologically interpretable insights into the pathophysiology of ME/CFS, highlighting its potential for clinical decision support and elucidating novel therapeutic targets.

Source: Yagin FH, Colak C, Al-Hashem F, Alzakari SA, Alhussan AA, Aghaei M. Leveraging Explainable Automated Machine Learning (AutoML) and Metabolomics for Robust Diagnosis and Pathophysiological Insights in Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS). Diagnostics (Basel). 2025 Oct 30;15(21):2755. doi: 10.3390/diagnostics15212755. PMID: 41226047. https://www.mdpi.com/2075-4418/15/21/2755 (Full text)

Circulating Levels of SMPDL3B Define Metabolic Endophenotypes and Subclinical Kidney Alterations in Myalgic Encephalomyelitis

Abstract:

Myalgic Encephalomyelitis (ME) is a complex, multisystem disorder with poorly understood pathophysiological mechanisms. SMPDL3B, a membrane-associated protein expressed in renal podocytes, is essential for lipid raft integrity and glomerular barrier function. We hypothesize that reduced membrane-bound SMPDL3B may contribute to podocyte dysfunction and impaired renal physiology in ME. To investigate this, we quantified soluble SMPDL3B in plasma and urine as a surrogate marker of membrane-bound SMPDL3B status and assessed renal clearance and plasma metabolomic profiles.
In a cross-sectional study of 56 ME patients and 16 matched healthy controls, ME patients exhibited significantly lower urine-to-plasma ratios of soluble SMPDL3B and reduced renal clearance, suggesting podocyte-related abnormalities. Plasma metabolomics revealed dysregulation of metabolites associated with renal impairment, including succinic acid, benzoic acid, phenyllactic acid, 1,5-anhydroglucitol, histidine, and citrate.
In ME patients, plasma SMPDL3B levels inversely correlated with 1,5-anhydroglucitol concentrations and renal clearance. Multivariable modeling identified the urine-to-plasma SMPDL3B ratio as an independent predictor of clearance. Female ME patients showed more pronounced SMPDL3B alterations, reduced clearance, and greater symptom severity. Non-linear associations between soluble SMPDL3B and lipid species further suggest systemic metabolic remodeling.
These findings support soluble SMPDL3B as a potential non-invasive biomarker of renal-podocyte involvement in ME, highlighting sex-specific differences that may inform future therapeutic strategies.
Source: Rostami-Afshari B, Elremaly W, McGregor NR, Huang KJK, Armstrong CW, Franco A, Godbout C, Elbakry M, Abdelli R, Moreau A. Circulating Levels of SMPDL3B Define Metabolic Endophenotypes and Subclinical Kidney Alterations in Myalgic Encephalomyelitis. International Journal of Molecular Sciences. 2025; 26(18):8882. https://doi.org/10.3390/ijms26188882 https://www.mdpi.com/1422-0067/26/18/8882 (Full text)

AI-driven multi-omics modeling of myalgic encephalomyelitis/chronic fatigue syndrome

Abstract:

Myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS) is a chronic illness with a multifactorial etiology and heterogeneous symptomatology, posing major challenges for diagnosis and treatment. Here we present BioMapAI, a supervised deep neural network trained on a 4-year, longitudinal, multi-omics dataset from 249 participants, which integrates gut metagenomics, plasma metabolomics, immune cell profiling, blood laboratory data and detailed clinical symptoms.

By simultaneously modeling these diverse data types to predict clinical severity, BioMapAI identifies disease- and symptom-specific biomarkers and classifies ME/CFS in both held-out and independent external cohorts. Using an explainable AI approach, we construct a unique connectivity map spanning the microbiome, immune system and plasma metabolome in health and ME/CFS adjusted for age, gender and additional clinical factors.

This map uncovers altered associations between microbial metabolism (for example, short-chain fatty acids, branched-chain amino acids, tryptophan, benzoate), plasma lipids and bile acids, and heightened inflammatory responses in mucosal and inflammatory T cell subsets (MAIT, γδT) secreting IFN-γ and GzA.

Overall, BioMapAI provides unprecedented systems-level insights into ME/CFS, refining existing hypotheses and hypothesizing unique mechanisms—specifically, how multi-omics dynamics are associated to the disease’s heterogeneous symptoms.

Source: Xiong, R., Aiken, E., Caldwell, R. et al. AI-driven multi-omics modeling of myalgic encephalomyelitis/chronic fatigue syndrome. Nat Med (2025). https://doi.org/10.1038/s41591-025-03788-3  https://www.nature.com/articles/s41591-025-03788-3

Untargeted analysis in post-COVID-19 patients reveals dysregulated lipid pathways two years after recovery

Abstract:

Introduction: Similar to what it has been reported with preceding viral epidemics (such as MERS, SARS, or influenza), SARS-CoV-2 infection is also affecting the human immunometabolism with long-term consequences. Even with underreporting, an accumulated of almost 650 million people have been infected and 620 million recovered since the start of the pandemic; therefore, the impact of these long-term consequences in the world population could be significant. Recently, the World Health Organization recognized the post-COVID syndrome as a new entity, and guidelines are being established to manage and treat this new condition. However, there is still uncertainty about the molecular mechanisms behind the large number of symptoms reported worldwide.

Aims and Methods: In this study we aimed to evaluate the clinical and lipidomic profiles (using non-targeted lipidomics) of recovered patients who had a mild and severe COVID-19 infection (acute phase, first epidemic wave); the assessment was made two years after the initial infection.

Results: Fatigue (59%) and musculoskeletal (50%) symptoms as the most relevant and persistent. Functional analyses revealed that sterols, bile acids, isoprenoids, and fatty esters were the predicted metabolic pathways affected in both COVID-19 and post-COVID-19 patients. Principal Component Analysis showed differences between study groups. Several species of phosphatidylcholines and sphingomyelins were identified and expressed in higher levels in post-COVID-19 patients compared to controls. The paired analysis (comparing patients with an active infection and 2 years after recovery) show 170 dysregulated features. The relationship of such metabolic dysregulations with the clinical symptoms, point to the importance of developing diagnostic and therapeuthic markers based on cell signaling pathways.

Source: López-Hernández Y, Oropeza-Valdez JJ, García Lopez DA, Borrego JC, Murgu M, Valdez J, López JA, Monárrez-Espino J. Untargeted analysis in post-COVID-19 patients reveals dysregulated lipid pathways two years after recovery. Front Mol Biosci. 2023 Mar 3;10:1100486. doi: 10.3389/fmolb.2023.1100486. PMID: 36936993; PMCID: PMC10022496. https://pmc.ncbi.nlm.nih.gov/articles/PMC10022496/ (Full text)

Advocating the role of trained immunity in the pathogenesis of ME/CFS: a mini review

Abstract:

Myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS) is a complex chronic disease of which the underlying (molecular) mechanisms are mostly unknown. An estimated 0.89% of the global population is affected by ME/CFS. Most patients experience a multitude of symptoms that severely affect their lives. These symptoms include post-exertional malaise, chronic fatigue, sleep disorder, impaired cognitive functions, flu-like symptoms, and chronic immune activation. Therapy focusses on symptom management, as there are no drugs available. Approximately 60% of patients develop ME/CFS following an acute infection.

Such a preceding infection may induce a state of trained immunity; defined as acquired, nonspecific, immunological memory of innate immune cells. Trained immune cells undergo long term epigenetic reprogramming, which leads to changes in chromatin accessibility, metabolism, and results in a hyperresponsive phenotype. Initially, trained immunity has only been demonstrated in peripheral blood monocytes and macrophages. However, more recent findings indicate that hematopoietic stem cells in the bone marrow are required for long-term persistence of trained immunity. While trained immunity is beneficial to combat infections, a disproportionate response may cause disease.

We hypothesize that pronounced hyperresponsiveness of innate immune cells to stimuli could account for the aberrant activation of various immune pathways, thereby contributing to the pathophysiology of ME/CFS. In this mini review, we elaborate on the concept of trained immunity as a factor involved in the pathogenesis of ME/CFS by presenting evidence from other post-infectious diseases with symptoms that closely resemble those of ME/CFS.

Source: Humer B, Dik WA, Versnel MA. Advocating the role of trained immunity in the pathogenesis of ME/CFS: a mini review. Front Immunol. 2025 Mar 25;16:1483764. doi: 10.3389/fimmu.2025.1483764. PMID: 40201181; PMCID: PMC11975576. https://pmc.ncbi.nlm.nih.gov/articles/PMC11975576/ (Full text)

The metabolic and physiologic impairments underlying long COVID associated exercise intolerance

Abstract:

Data from invasive CPET (iCPET) revealed long COVID patients have impaired systemic oxygen extraction (EO2), suggesting impaired mitochondrial ATP production. However, it remains uncertain whether the initial severity of SARS-CoV-2 infection has implications on EO2 and exercise capacity (VO2) nor has there been assessment of anerobic ATP generation in long COVID patients. iCPET was performed on 47 long COVID patients (i.e., full cohort; n = 8 with severe SARS-CoV-2 infection). ‘

In a subset of patients (i.e., metabolomic cohort; n = 26) metabolomics on venous and arterial blood samples during iCPET was performed. In the full cohort, long COVID patients exhibited reduced peak EO2 with reduced peak VO2 (90 ± 17% predicted) relative to cardiac output (118 ± 23% predicted). Peak VO2 [88% predicted (IQR 81% – 108%) vs. 70% predicted (IQR 64% – 89%); p = 0.02] and EO2 [0.59(IQR 0.53-0.62) vs. 0.53(IQR 0.50-0.48); p = 0.01) were lower in severe versus mild infection.

In the metabolomic cohort, 12 metabolites were significantly consumed, and 41 metabolites were significantly released (p-values < 0.05). Quantitative metabolomics demonstrated significant increases in inosine and succinate arteriovenous gradients during exercise. Peak VO2 was significantly correlated with peak venous succinate (r = 0.68; p = 0.0008) and peak venous lactate (r = 0.49; p = 0.0004). Peak EO2 and consequently peak VO2 impact long COVID patients in a severity dependent manner.

Exercise intolerance associated with long COVID is defined by impaired aerobic and anaerobic energy production. Peak venous succinate may serve as a potential biomarker in long COVID.

Source: Leitner BP, Joseph P, Quast AF, Ramirez MA, Heerdt PM, Villalobos JG, Singh I. The metabolic and physiologic impairments underlying long COVID associated exercise intolerance. Pulm Circ. 2024 Nov 13;14(4):e70009. doi: 10.1002/pul2.70009. PMID: 39544193; PMCID: PMC11560803. https://pmc.ncbi.nlm.nih.gov/articles/PMC11560803/ (Full text)

Cerebrospinal fluid metabolomics, lipidomics and serine pathway dysfunction in myalgic encephalomyelitis/chronic fatigue syndroome (ME/CFS)

Abstract:

We proposed that cerebrospinal fluid would provide objective evidence for disrupted brain metabolism in myalgic encephalomyelitis/chronic fatigue syndroome (ME/CFS). The concept of postexertional malaise (PEM) with disabling symptom exacerbation after limited exertion that does not respond to rest is a diagnostic criterion for ME/CFS. We proposed that submaximal exercise provocation would cause additional metabolic perturbations.

The metabolomic and lipidomic constituents of cerebrospinal fluid from separate nonexercise and postexercise cohorts of ME/CFS and sedentary control subjects were contrasted using targeted mass spectrometry (Biocrates) and frequentist multivariate general linear regression analysis with diagnosis, exercise, gender, age and body mass index as independent variables. ME/CFS diagnosis was associated with elevated serine but reduced 5-methyltetrahydrofolate (5MTHF).

One carbon pathways were disrupted. Methylation of glycine led to elevated sarcosine but further methylation to dimethylglycine and choline was decreased. Creatine and purine intermediates were elevated. Transaconitate from the tricarboxylic acid cycle was elevated in ME/CFS along with essential aromatic amino acids, lysine, purine, pyrimidine and microbiome metabolites. Serine is a precursor of phospholipids and sphingomyelins that were also elevated in ME/CFS. Exercise led to consumption of lipids in ME/CFS and controls while metabolites were consumed in ME/CFS but generated in controls.

The findings differ from prior hypometabolic findings in ME/CFS plasma. The novel findings generate new hypotheses regarding serine-folate-glycine one carbon and serine-phospholipid metabolism, elevation of end products of catabolic pathways, shifts in folate, thiamine and other vitamins with exercise, and changes in sphingomyelins that may indicate myelin and white matter dysfunction in ME/CFS.

Source: Baraniuk JN. Cerebrospinal fluid metabolomics, lipidomics and serine pathway dysfunction in myalgic encephalomyelitis/chronic fatigue syndroome (ME/CFS). Sci Rep. 2025 Mar 3;15(1):7381. doi: 10.1038/s41598-025-91324-1. PMID: 40025157. https://www.nature.com/articles/s41598-025-91324-1 (Full text)

The search for a blood-based biomarker for Myalgic Encephalomyelitis/ Chronic Fatigue Syndrome (ME/CFS): from biochemistry to electrophysiology

Abstract:

Background: Myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS) is a disease of unknown aetiology characterised by symptoms of post-exertional malaise (PEM) and fatigue leading to substantial impairment in functioning. Other key symptoms include cognitive impairment and unrefreshing sleep, with many experiencing pain. To date there is no complete understanding of the triggering pathomechanisms of disease, and no quantitative biomarker available with sufficient sensitivity, specificity, and adoptability to provide conclusive diagnosis. Clinicians thus eliminate differential diagnoses, and rely on subjective, unspecific, and disputed clinical diagnostic criteria-a process that often takes years with patients being misdiagnosed and receiving inappropriate and sometimes detrimental care. Without a quantitative biomarker, trivialisation, scepticism, marginalisation, and misunderstanding of ME/CFS continues despite the significant disability for many. One in four individuals are bed-bound for long periods of time, others have difficulties maintaining a job/attending school, incurring individual income losses of thousands, while few participate in social activities.

Main body: Recent studies have reported promising quantifiable differences in the biochemical and electrophysiological properties of blood cells, which separate ME/CFS and non-ME/CFS participants with high sensitivities and specificities-demonstrating potential development of an accessible and relatively non-invasive diagnostic biomarker. This includes profiling immune cells using Raman spectroscopy, measuring the electrical impedance of blood samples during hyperosmotic challenge using a nano-electronic assay, use of metabolomic assays, and certain techniques which assess mitochondrial dysfunction. However, for clinical application, the specificity of these biomarkers to ME/CFS needs to be explored in more disease controls, and their practicality/logistics considered. Differences in cytokine profiles in ME/CFS are also well documented, but finding a consistent, stable, and replicable cytokine profile may not be possible. Increasing evidence demonstrates acetylcholine receptor and transient receptor potential ion channel dysfunction in ME/CFS, though how these findings could translate to a diagnostic biomarker are yet to be explored.

Conclusion: Different biochemical and electrophysiological properties which differentiate ME/CFS have been identified across studies, holding promise as potential blood-based quantitative diagnostic biomarkers for ME/CFS. However, further research is required to determine their specificity to ME/CFS and adoptability for clinical use.

Source: Clarke KSP, Kingdon CC, Hughes MP, Lacerda EM, Lewis R, Kruchek EJ, Dorey RA, Labeed FH. The search for a blood-based biomarker for Myalgic Encephalomyelitis/ Chronic Fatigue Syndrome (ME/CFS): from biochemistry to electrophysiology. J Transl Med. 2025 Feb 4;23(1):149. doi: 10.1186/s12967-025-06146-6. PMID: 39905423.  https://translational-medicine.biomedcentral.com/articles/10.1186/s12967-025-06146-6 (Full text)

Exertional Exhaustion (Post-Exertional Malaise, PEM) Evaluated by the Effects of Exercise on Cerebrospinal Fluid Metabolomics–Lipidomics and Serine Pathway in Myalgic Encephalomyelitis/Chronic Fatigue Syndrome

Abstract

Post-exertional malaise (PEM) is a defining condition of myalgic encephalomyelitis (ME/CFS). The concept requires that a provocation causes disabling limitation of cognitive and functional effort (“fatigue”) that does not respond to rest. Cerebrospinal fluid was examined as a proxy for brain metabolite and lipid flux and to provide objective evidence of pathophysiological dysfunction. Two cohorts of ME/CFS and sedentary control subjects had lumbar punctures at baseline (non-exercise) or after submaximal exercise (post-exercise). Cerebrospinal fluid metabolites and lipids were quantified by targeted Biocrates mass spectrometry methods.
Significant differences between ME/CFS and control, non-exercise vs. post-exercise, and by gender were examined by multivariate general linear regression and Bayesian regression methods. Differences were found at baseline between ME/CFS and control groups indicating disease-related pathologies, and between non-exercise and post-exercise groups implicating PEM-related pathologies.
A new, novel finding was elevated serine and its derivatives sarcosine and phospholipids with a decrease in 5-methyltetrahydrofolate (5MTHF), which suggests general dysfunction of folate and one-carbon metabolism in ME/CFS. Exercise led to consumption of lipids in ME/CFS and controls while metabolites were consumed in ME/CFS but generated in controls. In general, the frequentist and Bayesian analyses generated complementary but not identical sets of analytes that matched the metabolic modules and pathway analysis. Cerebrospinal fluid is unique because it samples the choroid plexus, brain interstitial fluid, and cells of the brain parenchyma.
The quantitative outcomes were placed into the context of the cell danger response hypothesis to explain shifts in serine and phospholipid synthesis; folate and one-carbon metabolism that affect sarcosine, creatine, purines, and thymidylate; aromatic and anaplerotic amino acids; glucose, TCA cycle, trans-aconitate, and coenzyme A in energy metabolism; and vitamin activities that may be altered by exertion. The metabolic and phospholipid profiles suggest the additional hypothesis that white matter dysfunction may contribute to the cognitive dysfunction in ME/CFS.
Source: Baraniuk JN. Exertional Exhaustion (Post-Exertional Malaise, PEM) Evaluated by the Effects of Exercise on Cerebrospinal Fluid Metabolomics–Lipidomics and Serine Pathway in Myalgic Encephalomyelitis/Chronic Fatigue Syndrome. International Journal of Molecular Sciences. 2025; 26(3):1282. https://doi.org/10.3390/ijms26031282 https://www.mdpi.com/1422-0067/26/3/1282 (Full text)