Truace
Subscribe

Machine learning-aided serum N-glycomic signatures reveal potential biomarkers for acute ischemic stroke and minocycline treatment response

The absence of molecular-level tools for diagnosing and monitoring acute ischemic stroke (AIS) underscores the need for novel biomarkers. Serum N-glycosylation represents a promising but remains underexplored source of such markers in AIS. Using hydrophilic interaction liquid chromatography with fluorescence detection (HILIC-FLD), serum N-glycome profiling from healthy controls (HC), AIS patients who had not received prior treatment (AIS) and administered minocycline (A_M) were conducted. A computational pipelin…

By the Truvace desk2 min read1 sourcePeer-reviewedTRV-2026-1688· cite & verify
Machine learning-aided serum N-glycomic signatures reveal potential biomarkers for acute ischemic stroke and minocycline treatment response
"CT scan Rasmussen's encephalitis" by Mikhail Kalinin is licensed under CC BY-SA 3.0. To view a copy of this license, visit https://creativecommons.org/licenses/by-sa/3.0/.

In brief

Researchers profiled the serum N-glycome of healthy controls, untreated acute ischemic stroke patients, and patients given minocycline, using liquid chromatography with fluorescence detection. They then ran a pipeline of several machine-learning models and feature-selection algorithms on the glycomic data.

Stroke was linked to a glycan signature of increased branching and hypersialylation that was partly reversed after minocycline. If it holds up, a blood-based test could give clinicians a molecular way to diagnose stroke and monitor treatment, a tool the authors say is currently missing. The abstract reports no external validation, so real-world performance is unconfirmed.

Main points

  1. Stroke patients showed a serum N-glycome signature with more branching and hypersialylation.
  2. Minocycline treatment partially reversed this signature.
  3. The optimized random forest model, Glyco_RFF, uses six glycome features: GP14, GP32, GP33, GP34, S3 and S4.
  4. The model reached AUC values above 0.9 for stroke versus healthy controls and stroke versus minocycline-treated patients.
  5. Changes in GP34 and S4 over time correlated with NIHSS score reduction.

The gain

A random forest model using six serum glycan features distinguished stroke patients from healthy controls and from minocycline-treated patients, and two features tracked clinical improvement.

The rundown

The study compared three groups: healthy controls, acute ischemic stroke patients who had received no prior treatment, and patients given minocycline. Serum N-glycans were measured by hydrophilic interaction liquid chromatography with fluorescence detection.

The computational pipeline combined multiple machine-learning models with feature-selection algorithms. The result was a six-feature random forest model that separated both stroke from healthy samples and untreated stroke from minocycline-treated samples with high accuracy.

Two of the features, GP34 and S4, changed over time in step with clinical improvement as measured by NIHSS score reduction. The authors suggest these could serve as dynamic markers of treatment response, and they describe the model as a non-invasive tool for diagnosis and therapeutic monitoring.

Sources

  1. Peer-reviewedClinica Chimica Acta2026-10-08

The debate