Lorenzo De Gregorio
Association of the Capillary Metabolome with Cardiorespiratoy Fitness and Metabolic Response to Acute Exercise.
Rel. Gabriella Olmo. Politecnico di Torino, Corso di laurea magistrale in Data Science And Engineering, 2026
|
Preview |
PDF (Tesi_di_laurea)
- Tesi
Licenza: Creative Commons Attribution Non-commercial No Derivatives. Download (3MB) | Preview |
Abstract
Acute exercise metabolomics is commonly based on a single pre-exercise measurement, which limits the ability to determine whether post-exercise changes exceed short-term resting variability. This thesis investigated whether repeated-baseline capillary blood metabolomics can improve the interpretation of acute metabolic responses to maximal exercise and whether baseline metabolomic information contributes to the prediction of cardiorespiratory fitness. A total of 100 healthy adults completed maximal cardiopulmonary exercise testing. Capillary blood samples were collected at two resting baseline timepoints and immediately after exercise. Targeted metabolomics data were filtered to retain 52 metabolites, including amino acids, amino acid-related metabolites, fatty acids, and lactate. Baseline reproducibility was assessed using within-subject variability, correlations and exploratory dimensionality reduction.
Acute exercise responses were evaluated through principal component analysis, conventional single-baseline comparisons and a signal-to-noise ratio framework, in which exercise reactivity was interpreted relative to baseline-derived methodological noise. Participant-level determinant analyses examined the associations of physiological and experimental characteristics with noise-adjusted metabolite responses. In addition, baseline metabolomic and participant-level features were used to predict VO₂peak through regularized regression and nested cross-validation. The repeated baseline samples showed strong overall agreement, with a global correlation of R = 0.996 and a median metabolite-level correlation of R = 0.830. Principal component analysis showed a clear separation between resting and post-exercise samples, indicating a broad exercise-induced shift in the capillary metabolome. Forty of the 52 retained metabolites were significantly altered after exercise, with lactate showing the strongest response. However, signal-to-noise stratification showed that response robustness varied substantially across metabolites, with 46 metabolites displaying an average post-exercise response larger than baseline-derived variability. Exhaustion status and fasting status were the main determinants of inter-individual response variability. Participants classified as potentially not exhausted showed attenuated amino acid-related responses, whereas fasting status was associated with a smaller subset of metabolite responses. The prediction analysis showed that VO₂peak was estimated with moderate accuracy from participant-level clinical features (R2 = 0.606; MAE = 4.322 ml min-1 kg-1), whereas baseline metabolomic features provided limited standalone and incremental predictive value. Overall, repeated-baseline capillary metabolomics provides a useful framework for interpreting acute exercise responses relative to short-term resting variability. The main contribution of this thesis is methodological: repeated baseline sampling supports a noise-aware interpretation of metabolite changes and helps distinguish broad group-level perturbations from responses that more clearly exceed baseline-derived methodological noise.
Relatori
Anno Accademico
Tipo di pubblicazione
Numero di pagine
Corso di laurea
Classe di laurea
Aziende collaboratrici
URI
![]() |
Modifica (riservato agli operatori) |
