Related Experiment Video
Updated: Jan 11, 2026
07:08
A High Throughput, Multiplexed and Targeted Proteomic CSF Assay to Quantify Neurodegenerative Biomarkers and Apolipoprotein E Isoforms Status
Published on: October 20, 2016
8.2K
Accurate diagnosis of hemoglobinopathies with machine learning based on high-throughput proteomics
Shaodong Wei1,2, Annelaura Bach Nielsen2,3, Jens Helby1,4
1Department of Hematology, Danish Red Blood Cell Center Copenhagen University Hospital-Rigshospitalet Copenhagen Denmark.
Hemasphere
|November 14, 2025
Summary
A new proteomics and machine learning method accurately diagnoses hemoglobinopathies like sickle cell disease and thalassemias. This approach offers a faster, scalable, and cost-effective alternative to current diagnostic tools.
Area of Science:
- Biochemistry
- Genetics
- Computational Biology
Background:
- Hemoglobinopathies present a significant global health challenge, particularly in endemic areas.
- Existing diagnostic methods (HPLC, electrophoresis, genetic testing) have limitations in speed, cost, and comprehensive variant detection.
Purpose of the Study:
- To develop and validate a novel diagnostic framework for hemoglobinopathies using high-throughput proteomics and machine learning.
- To overcome the limitations of current diagnostic techniques for structural hemoglobin variants and beta-thalassemia trait.
Main Methods:
- Analysis of tryptic peptides from red blood cells, whole blood, and plasma using data-independent acquisition mass spectrometry.
- Development and validation of random forest classifiers to identify structural hemoglobin variants and beta-thalassemia trait.
- Evaluation of model performance using Monte Carlo cross-validation on development and validation cohorts.
Main Results:
- The proteomics-based classifier achieved an AUC of 1.000 and 99.9% accuracy for structural variants in the validation cohort.
- For beta-thalassemia trait, the mean AUC was 1.000 and prediction accuracy reached 96.9% in the validation cohort.
- A single peptide decision tree demonstrated 92% accuracy for beta-thalassemia trait diagnosis.
Conclusions:
- High-throughput proteomics combined with machine learning provides a rapid, scalable, and potentially cost-effective diagnostic solution for hemoglobinopathies.
- This novel approach minimizes sample preparation and manual interpretation, offering precise classification of hereditary anemias.
- The framework shows promise for routine clinical application in diagnosing and managing hemoglobinopathies.
