Aristotle University Of Thessaloniki (AUTH)
Big data analytics for exposomics: ionizing radiation, data harvesting, and environmental modulation
Abstract
dc:descriptionThis doctoral dissertation develops an integrated computational framework for exposomics, the study of how environmental exposures accumulate and interact over time to influence biological function and human health. The research focuses on two physical environmental factors, ultraviolet and ionizing radiation, which are examined not as isolated risk variables but as dynamic, structured signals acting across molecular, biological, and population levels. The theoretical and methodological foundation of the dissertation is systems radiobiology. In contrast to classical approaches that focus on individual endpoints, the proposed framework uses gene expression, biological pathway activity, and molecular network behavior to identify subtle, nonlinear, and temporally distributed responses to low-dose radiation. This approach requires analytical infrastructures capable of handling large, heterogeneous, and high-dimensional datasets, with particular emphasis on reproducibility and interpretability. The first major contribution investigates ultraviolet radiation as a seasonal environmental modulator of COVID-19 dynamics. A multivariable dataset was developed by integrating epidemiological, environmental, socioeconomic, and policy-related indicators across multiple countries and time periods. Through time-series analysis, normalization, non-parametric association measures, lagged correlations, and comparisons between the two hemispheres, consistent seasonal and hemispheric patterns were identified. These findings are compatible with the hypothesis that ultraviolet radiation acts as one of several environmental factors shaping epidemic dynamics. The second contribution is the development of BRIDE v2, a curated gene collection designed to support the study of low-dose ionizing-radiation responses in the mammalian brain. The resource integrates dispersed transcriptomic evidence and is organized according to recurrence of findings, biological coherence, and relevance to critical processes, including DNA-damage response, oxidative stress, immune signaling, mitochondrial function, and neuroinflammation. The third contribution is the Composite Morbidity Index (CMI), an interpretable composite indicator that integrates normalized epidemiological and environmental features. The index was developed using XGBoost and SHAP so that the contribution of each variable remains transparent and can be compared across regions, time periods, and seasonal conditions. Overall, the dissertation brings these contributions together within a multi-layer analytics and decision-support framework connecting environmental exposure, biological response, and population-level morbidity. Exposure signals, molecular responses, and epidemiological indicators are maintained as distinct but interacting analytical layers rather than being collapsed into a single opaque model. A central principle throughout the dissertation is that interpretability is not an optional technical feature but a fundamental scientific requirement. In complex exposure environments, meaningful analysis must extend beyond correlation and prediction toward reasoning that is transparent, biologically grounded, reproducible, and useful for decision-making in public health, environmental regulation, occupational health, and biomedical research.
Degree
thesis:*- Grantor dc:publisher
- Aristotle University Of Thessaloniki (AUTH)
- Year dc:date
- 2026
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Karapiperis, Christos
- Καραπιπέρης, Χρήστος
Subjects
dc:subject × 14- Εκθεσίωμα
- Συστημική Ακτινοβιολογία
- Αναλύσεις Μεγάλων Δεδομένων
- Ερμηνεύσιμη τεχνητή νοημοσύνη
- Exposomics
- Systems Radiobiology
- Big data analytics
- Explainable artificial intelligence
- Άλλες Επιστήμες Μηχανικού και Τεχνολογίες
- Επιστήμες Μηχανικού και Τεχνολογία
- Μηχανική, διεπιστημονική προσέγγιση
- Other Engineering and Technologies
- Engineering and Technology
- Engineering, interdisciplinary
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/62342