Rare Disease Biomarkers & Computational Phenotyping

national
bioinformatics
rare-diseases
Age-adjusted reference intervals for CSF amino acids and serum copper in large paediatric cohorts, in long-standing collaboration with Dr. Rafael Artuch at Hospital Sant Joan de Déu.
Keywords

rare diseases, cerebrospinal fluid, CSF, amino acids, copper, paediatric, reference intervals, neurometabolic, epilepsy, biogenic amines, Sant Joan de Déu, UPC, Ministerio de Ciencia

Funding Ministerio de Ciencia e Innovación (AEI)
Programme Proyectos de Generación de Conocimiento / Proyectos I+D+i
Status Active
Clinical partner Hospital Sant Joan de Déu, Barcelona
Key collaborator Dr. Rafael Artuch Iriberri
Doctoral student Helena Rodríguez-González (co-advised by Perera-Lluna and Artuch)
B2SLab PI Alexandre Perera Lluna

Context: rare neurometabolic disease and the biomarker gap

Inherited neurometabolic disorders are individually rare but collectively significant — they affect thousands of children across Europe, often presenting in infancy or early childhood with seizures, developmental delay, movement disorders, or unexplained metabolic crises. Diagnosing them requires precise biochemical measurement: amino acids, biogenic amines, organic acids, trace elements, and their metabolic intermediates in cerebrospinal fluid (CSF) and blood.

The clinical problem is deceptively simple: to know whether a patient’s result is abnormal, you need to know what normal looks like — for a child of that age, in that biological context. In adults, this is hard enough. In children, it is substantially harder: concentrations of neurochemicals in CSF shift dramatically across the first years of life, driven by brain maturation, myelination, nutritional requirements, and the changing activity of metabolic pathways. Reference intervals derived from a single age group, from a small convenience sample, or from adults translated downward are systematically misleading.

B2SLab addresses this problem at its computational root — building statistically rigorous, age-continuous reference intervals for paediatric biomarkers using the large, carefully collected clinical datasets that only a centre of excellence like Hospital Sant Joan de Déu can accumulate.


A long-standing collaboration with Dr. Rafael Artuch

The intellectual and clinical engine of this research line is a sustained collaboration between B2SLab and Dr. Rafael Artuch Iriberri, Head of Clinical Biochemistry and Neurometabolism at Hospital Sant Joan de Déu (SJD) in Barcelona.

Dr. Artuch is one of Europe’s foremost experts in the biochemical diagnosis of neurometabolic diseases. Under his leadership, the SJD Clinical Biochemistry Service has built one of the largest and most carefully annotated collections of paediatric CSF and blood samples in the world, accumulated over decades of diagnostic activity for children referred with neurological disease from across Catalonia and beyond.

Tip

What this collaboration provides

  • Access to uniquely large paediatric cohorts — hundreds to thousands of samples — with the clinical annotation needed to distinguish true reference populations from disease-affected cases
  • Deep expertise in neurometabolic disease biology that guides which biomarkers matter, which confounders must be modelled, and how results translate into clinical decision-making
  • A direct pathway from statistical methodology to validated clinical tools deployed in real diagnostic workflows at one of Europe’s leading paediatric centres

This is a collaboration built on genuine mutual need: the clinical datasets at SJD require sophisticated statistical and computational methods that the clinical laboratory cannot provide alone; the computational methods require clinical data and expert interpretation that B2SLab cannot generate independently. The result is a programme in which methodology and biology advance together.


CSF amino acids: reference intervals for a diagnostic gold standard

Why CSF amino acids matter

Cerebrospinal fluid amino acid profiling is the diagnostic gold standard for a range of inherited neurometabolic conditions: disorders of amino acid transport across the blood-brain barrier, neurotransmitter synthesis defects, urea cycle disorders with CNS involvement, and the nutritional consequences of severe gastrointestinal disease. It is also increasingly used to monitor treatment response in conditions such as phenylketonuria and non-ketotic hyperglycinaemia.

The challenge is interpretive: CSF is collected by lumbar puncture, which limits sampling to patients with clinical indication. The “healthy reference population” from which reference intervals should ideally be derived is therefore never truly healthy — it consists of children who underwent lumbar puncture for a clinical reason but were found to have no metabolic disease. Managing this systematic selection bias, alongside the strong age-dependence of amino acid concentrations, is the central methodological problem.

Study: CSF amino acid profiling in a large paediatric cohort

Note

Neurobiology of Disease, 2025

Working with Dr. Artuch’s group at SJD, B2SLab analysed 410 CSF samples collected from neuropaediatric patients between 2018 and 2023 using ultrahigh-performance liquid chromatography with tandem mass spectrometry (UPLC-MS/MS).

The cohort comprised: - 201 children with epilepsy — spanning focal, generalised, and unclassified seizure types - 209 disease controls — children with other neurological presentations but no metabolic disease diagnosis

Key findings:

  • CSF amino acid concentrations are highest in the neonatal period and decline through early childhood, with some amino acids rising again in later childhood — confirming that age-continuous reference intervals are essential and that age-stratified binning introduces systematic error
  • More than 90% of children with epilepsy had CSF amino acid levels within normal age-adjusted ranges, establishing that epilepsy per se does not produce generalised amino acid abnormalities — an important baseline for clinical interpretation
  • Antiseizure medications alter CSF amino acid profiles: valproate and GABA-receptor-acting drugs elevated glutamine; vigabatrin elevated ornithine, leucine, and isoleucine — drug effects that must be distinguished from disease effects when interpreting clinical results
  • A subset of patients with generalised epilepsy showed elevated essential amino acids, consistent with blood-brain barrier disruption — a finding that opens a new research direction into the metabolic consequences of severe seizures

The study produced the largest set of age-adjusted CSF amino acid reference intervals for children published to date. These are available through an open online calculator for use by clinical laboratories.

Rodriguez-Gonzalez H, Nou-Fontanet L, Ormazabal A, Casado M, Arias Á, Pias-Peleteiro L, Gutierrez A, Perera-Lluna A, Garcia-Cazorla A, Fons C, Artuch R. Neurobiology of Disease, 2025. doi:10.1016/j.nbd.2025.107098


Serum copper: when inflammation confounds the measurement

The diagnostic importance of copper

Copper is an essential trace element — cofactor for key enzymes in electron transport, antioxidant defence, iron metabolism, and neuropeptide synthesis. Both deficiency (Menkes disease, nutritional deprivation, malabsorption) and excess (Wilson disease, toxic exposure) cause serious disease, with neurological consequences prominent in both directions. In children, the biochemical diagnosis and monitoring of copper metabolism disorders depends critically on reliable reference intervals for serum copper and its carrier protein ceruloplasmin.

The confounding effect of inflammation

A complication rarely addressed in existing reference interval studies is that copper is an acute-phase reactant: inflammation drives increased hepatic synthesis of ceruloplasmin, the main copper-carrying protein in blood, raising measured serum copper independently of true copper status. A child with an ongoing infection, autoimmune condition, or chronic inflammatory disease will appear copper-replete on a standard reference range comparison even if their true functional copper status is compromised.

Study: age- and inflammation-adjusted copper reference intervals

Note

Clinical Nutrition, 2026

Building on the continuous reference interval methodology developed for CSF amino acids, B2SLab and SJD analysed 4,368 paediatric serum samples — one of the largest paediatric biomarker datasets of this kind — to construct the first age- and inflammation-adjusted reference intervals for serum copper in children.

Approach:

  1. Modelled copper concentration as a continuous function of age using nonlinear regression, capturing the full developmental trajectory from neonates through adolescents
  2. Quantified the contribution of inflammation by measuring three established inflammatory markers — erythrocyte sedimentation rate (ESR), fibrinogen, and C-reactive protein (CRP)
  3. Built a composite inflammation score using partial least squares regression across the three markers, then incorporated this score as an additional covariate in the reference interval model

Key finding: Inflammation elevated measured serum copper by approximately 24% on average. Children with active inflammation assessed against standard copper reference intervals will systematically appear copper-replete — potentially masking true deficiency.

The result is a proof-of-concept model for clinically adjusted copper interpretation: a tool that allows a clinician to enter a child’s age, measured copper, and inflammatory marker values and receive an inflammation-corrected interpretation of whether the result falls within the expected range.

Rodriguez-Gonzalez H, Arias A, Poyatos E, et al. Clinical Nutrition, 2026. doi:10.1016/j.clnu.2026.106586


Helena Rodríguez-González: doctoral thesis

Both research papers above are primary outputs of the doctoral thesis of Helena Rodríguez-González, jointly supervised by Prof. Alexandre Perera-Lluna (UPC / B2SLab) and Dr. Rafael Artuch Iriberri (Hospital Sant Joan de Déu).

Helena’s thesis sits at the boundary between clinical biochemistry and computational methods. The unifying thread across her work is the problem of clinical reference interval computation for paediatric biomarkers: how to derive, validate, and communicate normal ranges from complex observational data, accounting for age-dependence, biological confounders, sampling bias, and the particular challenges of rare disease populations.

The co-supervision model reflects the nature of the problem: the statistical and computational methodology is developed at B2SLab; the clinical datasets, biological interpretation, and validation against diagnostic outcomes are provided by Dr. Artuch’s team at SJD. The collaboration ensures that methodological advances are grounded in real clinical need and that clinical datasets are analysed with the rigour that complex observational data demands.


Broader context

This research line connects directly to B2SLab’s wider rare disease work:

  • Share4Rare — the European patient data platform for rare diseases, where B2SLab designed the HPO-based phenotype standardisation infrastructure used to characterise patients across 50+ countries
  • Metabolomics for Metabolic Disease — the complementary line using untargeted LC-MS to discover novel biomarkers for common metabolic diseases