Related Experiment Video
Updated: Jun 21, 2025
09:47
Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches
Published on: December 15, 2023
1.0K
Decoding depression: a comprehensive multi-cohort exploration of blood DNA methylation using machine learning and
Aleksandr V Sokolov1, Helgi B Schiöth2
1Department of Surgical Sciences, Functional Pharmacology and Neuroscience, Uppsala University, Uppsala, Sweden.
Translational Psychiatry
|July 15, 2024
Summary
This study explored blood DNA methylation for depression diagnosis. Machine learning models showed potential, with Random Forest achieving high accuracy, highlighting methylation signatures as possible biomarkers.
Area of Science:
- Neuroscience and Genomics
- Computational Biology and Bioinformatics
Background:
- Current depression diagnosis lacks objective laboratory biomarkers, relying solely on psychiatric evaluations.
- Investigating blood DNA methylation signatures offers a potential avenue for objective depression assessment.
Purpose of the Study:
- To assess the stability and diagnostic potential of blood DNA methylation signatures for depression across diverse populations.
- To evaluate various machine learning and deep learning strategies for depression classification using methylation data.
Main Methods:
- Mega-analysis and meta-analysis of DNA methylation data from 1942 individuals across 8 cohorts.
- Evaluation of 12 machine learning/deep learning models with both biased and unbiased feature selection.
- Cross-validation and hold-out testing on batch-level processed and harmonized data.
Main Results:
- Identified 1987 CpG sites associated with depression, linked to axon guidance and immune pathways.
- Random Forest classifiers achieved highest performance (AUC 0.73-0.76) on batch-level data.
- Models with pre-selected features significantly improved performance, with some reaching AUCs up to 0.91.
Conclusions:
- Blood DNA methylation signatures show promise as biomarkers for depression.
- Machine learning, particularly Random Forest, is effective for analyzing methylation data.
- Methodological considerations, including feature selection and data harmonization, are critical for reliable DNA methylation profiling.
