Federated Learning for multi-omics: a performance evaluation in Parkinson's disease.

The wide-scale application of Artificial Intelligence and computationally intensive analytical approaches in the biomedical and clinical domain is largely restricted by access to sufficient training data. This data scarcity exists due to the siloed nature of biomedical and clinical institutions, mandated by patient privacy policies in the health system or government legislation. The authors study the feasibility of applying Federated Learning (FL), a machine learning approach that facilitates sample-private, collaborative model training, to multi-omics Parkinson's disease prediction. Additionally, the authors describe the performance characteristics that FL exhibits to better understand the opportunities and challenges in applying such methods in the broader biomedical research community. Federated Learning approach will enable more productive and resource-efficient collaborations across research institutions; it provides access to high-quality datasets and enhances deeper analysis and, ultimately, the nascence of large-scale precision medicine studies.