# Assessing the Reproducibility of Machine-learning-based Biomarker Discovery in Parkinson's Disease

arXiv: https://arxiv.org/abs/2304.03239 · HTML: https://mlsanity.com/paper/2304.03239/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 0 of its 68 references that are in the map.
- First posted on arXiv 6 Apr 2023.

## Built on by (most cited first, 1 of 1)

- [Stabilizing Machine Learning for Reproducible and Explainable Results: A Novel Validation Approach to Subject-Specific Insights](https://mlsanity.com/paper/2412.16199/) (2024)

## Abstract (arXiv)

Genome-Wide Association Studies (GWAS) help identify genetic variations in people with diseases such as Parkinson's disease (PD), which are less common in those without the disease. Thus, GWAS data can be used to identify genetic variations associated with the disease. Feature selection and machine learning approaches can be used to analyze GWAS data and identify potential disease biomarkers. However, GWAS studies have technical variations that affect the reproducibility of identified biomarkers, such as differences in genotyping platforms and selection criteria for individuals to be genotyped. To address this issue, we collected five GWAS datasets from the database of Genotypes and Phenotypes (dbGaP) and explored several data integration strategies. We evaluated the agreement among different strategies in terms of the Single Nucleotide Polymorphisms (SNPs) that were identified as potential PD biomarkers. Our results showed a low concordance of biomarkers discovered using different datasets or integration strategies. However, we identified fifty SNPs that were identified at least twice, which could potentially serve as novel PD biomarkers. These SNPs are indirectly linked to PD in the literature but have not been directly associated with PD before. These findings open up new potential avenues of investigation.