Aziz, N. A. | bioRxiv Science

bioRxiv2020

Swarm Learning as a privacy-preserving machine learning approach for disease classification

Identification of patients with life-threatening diseases including leukemias or infections such as tuberculosis and COVID-19 is an important goal of precision medicine. We recently illustrated that leukemia patients are identified by machine learning (ML) based on their blood transcriptomes. However, there is an increasing divide between what is technically possible and what is allowed because of privacy legislation. To facilitate integration of any omics data from any data owner world-wide without violating privacy laws, we here introduce Swarm Learning (SL), a decentralized machine learning approach uniting edge computing, blockchain-based peer-to-peer networking and coordination as well as privacy protection without the need for a central coordinator thereby going beyond federated learning. Using more than 14,000 blood transcriptomes derived from over 100 individual studies with non-uniform distribution of cases and controls and significant study biases, we illustrate the feasibility of SL to develop disease classifiers based on distributed data for COVID-19, tuberculosis or leukemias that outperform those developed at individual sites. Still, SL completely protects local privacy regulations by design. We propose this approach to noticeably accelerate the introduction of precision medicine.

bioinformatics↗

bioRxiv2017

Transfer entropy as a tool for inferring causality from observational studies in epidemiology

Recently Wieners causality theorem, which states that one variable could be regarded as the cause of another if the ability to predict the future of the second variable is enhanced by implementing information about the preceding values of the first variable, was linked to information theory through the development of a novel metric called transfer entropy. Intuitively, transfer entropy can be conceptualized as a model-free measure of directed information flow from one variable to another. In contrast, directionality of information flow is not reflected in traditional measures of association which are completely symmetric by design. Although information theoretic approaches have been applied before in epidemiology, their value for inferring causality from observational studies is still unknown. Therefore, in the present study we use a set of simulation experiments, reflecting the most classical and widely used epidemiological observational study design, to validate the application of transfer entropy in epidemiological research. Moreover, we illustrate the practical applicability of this information theoretic approach to real-world epidemiological data by demonstrating that transfer entropy is able to extract the correct direction of information flow from longitudinal data concerning two well-known associations, i.e. that between smoking and lung cancer and that between obesity and diabetes risk. In conclusion, our results provide proof-of-concept that the recently developed transfer entropy method could be a welcome addition to the epidemiological armamentarium, especially to dissect those situations in which there is a well-described association between two variables but no clear-cut inclination as to the directionality of the association.

epidemiology↗