Machine learning applied to higher order functional representations of omics data reveals biological pathways associated with Parkinson‘s Disease

Abstract

Despite the increasing prevalence of Parkinson’s Disease (PD) and research efforts to understand its underlying molecular pathogenesis, early diagnosis of PD remains a challenge. Machine learning analysis of omics data is a promising non-invasive approach to finding molecular fingerprints associated with PD that may enable an early and accurate diagnosis. We applied several machine learning classification methods to public omics data from PD case/control studies. We used aggregation statistics and Pathifier’s pathway deregulation scores to generate higher order pathway-level features and compared the models’ performance and most relevant predictive features with individual feature level predictors. The resulting diagnostic models from individual features and Pathifier’s pathway deregulation scores achieve significant Area Under the Curve (AUC, a receiver operating characteristic curve) scores for both cross-validation and external testing. Furthermore, we identify plausible biological pathway associations. We have successfully built machine learning models at pathway-level and single-feature level to study omics data for PD diagnosis. Furthermore, we show that pathway deregulation scores can serve as robust and biologically interpretable predictors for PD.

Elisa G. de Lope

PhD student in Biomedical Data Science

My research interests include statistics, data mining, -omics, and drug discovery.