Identification of Glycolysis-Related Diagnostic Biomarkers for Amyotrophic Lateral Sclerosis Using Machine Learning.

Abstract

Background: Glycometabolism has been implicated in the pathogenesis of amyotrophic lateral sclerosis (ALS), yet the precise molecular mechanisms underlying this association remain poorly understood. The identification of reliable biomarkers for ALS diagnosis represents a critical unmet need in clinical practice, as early detection and intervention could significantly improve patient outcomes.

Methods: We employed a comprehensive analytical approach combining two-sample Mendelian randomization analysis to investigate the causal relationship between blood glucose levels and ALS. Additionally, we integrated differential expression analysis, multiple machine learning algorithms, and correlation analyses to identify potential diagnostic biomarkers for ALS. The machine learning framework utilized gradient boosting tree methodology to construct predictive models, with performance evaluation conducted through cross-validation procedures.

Results: Mendelian randomization analysis demonstrated a significant negative causal relationship between blood glucose levels and ALS risk. Through bioinformatic analysis and machine learning approaches, we successfully identified candidate genes and constructed a high-performance predictive model using gradient boosting tree methodology, achieving an average area under the curve (AUC) of 0.8782 in cross-validation. Validation studies utilizing both bulk and single-cell RNA sequencing datasets revealed that COL5A1 and VCAN genes play significant roles in ALS pathogenesis, likely through their involvement in glycolytic pathways.

Conclusions: Our findings provide novel insights into the molecular mechanisms linking glycometabolism and ALS, while identifying potential diagnostic biomarkers for the disease. The identified genes, COL5A1 and VCAN, represent promising targets for further investigation in ALS pathogenesis. However, the clinical translation of these findings requires validation through additional datasets and prospective clinical trials to establish their diagnostic utility and therapeutic potential.