Blood-based DNA methylation markers for autism spectrum disorder identification using machine learning.

Area of Science:

• Epigenetics
• Computational Biology
• Neurodevelopmental Disorders

Background:

• Autism spectrum disorder (ASD) lacks objective biomarkers for early diagnosis.
• DNA methylation is a potential epigenetic marker for ASD.
• Few studies have explored genome-wide DNA methylation in whole blood for ASD diagnosis in children.

Purpose of the Study:

• To investigate the feasibility of using genome-wide DNA methylation profiles from peripheral blood for ASD diagnosis in school-aged children.
• To apply machine learning algorithms to classify ASD based on DNA methylation patterns.
• To identify potential epigenetic biomarkers for ASD.

Main Methods:

• Analysis of genome-wide DNA methylation data from 52 children with ASD and 48 typically developing (TD) controls.
• Identification of differentially methylated positions (DMPs) and feature selection using Support Vector Machine-Recursive Feature Elimination with Cross-Validation (SVM-RFECV).
• Development and evaluation of classification models including Random Forest (RF), Extreme Gradient Boosting (XGBoost), and Decision Tree (DT).

Main Results:

• 138 DMPs were identified between ASD and TD children.
• Eleven CpG sites selected by SVM-RFECV were used for model construction.
• RF and XGBoost models achieved 75% accuracy, while the DT model reached 70% accuracy. Functional annotation revealed enrichment in cell adhesion and immune-related pathways.

Conclusions:

• Peripheral blood DNA methylation data combined with machine learning can distinguish children with ASD.
• This study provides methodological insights into integrating epigenetic and computational approaches for ASD biomarker discovery.
• Further research with larger sample sizes is needed to improve accuracy and validate findings.