68994 research outputs found
Sort by
AI-assisted methodology for robust digital measurements by Raman spectroscopy: Quantification of inorganic pollutants in water
No full textRaman spectroscopy is a versatile analytical tool, yet it often struggles with low sensitivity, hardware noise, and environmental interference. To address these limitations, this study presents an automated, Artificial Intelligence (AI)-assisted methodology to convert noisy optical signals into robust digital measurements. The process involves acquiring high-dimensional, noisy spectral data from analyte solutions. A grid search across various algorithms identifies the optimal pre-processing pipeline to minimize noise variance and ensure metrological repeatability. Instead of relying on raw sensor feeds, the method fits a Gaussian curve combined with a polynomial baseline to the data, extracting precise measurements from the peak of this mathematical model. Supported by AI, the method successfully separates multiple optical signals and their shifts originating from interactions among analytes, proving itself capable to compensate also for possible hardware misalignment and thermal drift. As such, it can be used to quantify the concentration of selected inorganic pollutants in a mixture of analytes. The primary application addressed in this work is quantifying inorganic pollutants in water, to enable in situ analysis without continuous expert supervision. Tests on binary and ternary mixtures of inorganic pollutants in pure water demonstrated that the Mean Absolute Percentage Error (MAPE) for nitrate was consistently below 10% in the concentration range between 0 mg/L to more than 15 000 mg/L, dropping to under 5% for concentrations exceeding 1000 mg/L. For concentrations below 1000 mg/L, the Mean Absolute Error (MAE) values were 67 mg/L for nitrate, 1475 mg/L for sulfate, and 736 mg/L for nitrite, respectively
Advancing Human-Machine Symbiosis: Data Fusion Approaches for Analyzing Driver Behavior and Vehicle Interactions
No full textQuesta ricerca ha esplorato la progettazione e lo sviluppo di un'interfaccia uomo-macchina (HMI) simbiotica integrata con un avanzato sistema di monitoraggio del conducente (DMS), con l'obiettivo di migliorare la sicurezza, il coinvolgimento e l'esperienza complessiva di guida. Concentrandosi su interazioni adattive e predittive, lo studio mira a promuovere la ricerca su una relazione simbiotica tra conducente e veicolo. Il sistema è in grado di monitorare attivamente gli stati del conducente e fornire feedback in tempo reale per migliorare attenzione, comportamento e consapevolezza situazionale.
La ricerca ha indagato come le emozioni e i comportamenti del conducente si correlino con cambiamenti nell’attenzione e nelle prestazioni. A tal fine, sono state impiegate tecniche di fusione dei dati per integrare input multimodali, tra cui dati telemetrici, fisiologici e osservazionali. Le fasi sperimentali sono state condotte utilizzando un simulatore di guida dotato di telecamere e software di apprendimento automatico. Questi strumenti hanno permesso il rilevamento degli stati del conducente, come l’orientamento della testa e l’analisi delle emozioni attraverso il riconoscimento delle espressioni facciali basato sul modello emozionale di Ekman, facilitando l’analisi dei livelli di attenzione e delle risposte emotive durante eventi di distrazione.
I risultati principali evidenziano l’importanza del monitoraggio in tempo reale e dei meccanismi di feedback adattivi. Il modello di base multimodale CNN ha raggiunto prestazioni elevate, con un’accuratezza multiclasse vicina al 90% e un’accuratezza binaria superiore al 93%.
Questa ricerca contribuisce all’avanzamento della simbiosi uomo-macchina nel settore automobilistico, offrendo un framework per comprendere i comportamenti del conducente, migliorare l’interazione con il veicolo e aumentare la sicurezza stradale. I lavori futuri si concentreranno sul perfezionamento delle metodologie, sull’ottimizzazione degli strumenti di analisi e sulla validazione dei risultati in ambienti reali.This research explored the design and development of a symbiotic Human-Machine Interface (HMI) integrated with an advanced Driver Monitoring System (DMS) to enhance driver safety, engagement, and overall driving experience. By focusing on adaptive and predictive interactions, the study aims to advance research on a symbiotic relationship between the driver and the vehicle. The system can actively monitor driver states and provide real-time feedback to improve attention, behavior, and situational awareness.
The research investigated how driver’s emotions and behaviors correlate with changes in attention and performance. To this end, the study uses data fusion techniques to integrate multimodal inputs, including telemetric, physiological and observational data. Experimental phases were conducted using a driving simulator equipped with cameras and machine learning software. These tools enabled the detection of driver states such as head orientation and emotional analysis through facial expression recognition based on Ekman's emotion framework, facilitating the analysis of attention levels and emotional responses during distraction events.
Key findings underline the importance of real-time monitoring and adaptive feedback mechanisms. Results show that the implemented baseline multimodal CNN model achieved a high degree of performance, with multiclass accuracy nearing 90% and binary classification accuracy exceeding 93%.
This research contributes to advancing Human-Machine Symbiosis in the automotive sector by offering a framework for understanding driver behaviors, enhancing vehicle interaction, and improving road safety. Future work will focus on refining methodologies, optimizing analysis tools, and validating findings in real-world environments
Environmental sustainability assessment of remediation alternatives for highly contaminated marine sediments
No full textThis study compares the environmental sustainability of five alternatives for the remediation of marine sediments of one of the most polluted coastal sites in Europe (Bagnoli-Coroglio bay, Mediterranean Sea), using the Life Cycle Assessment (LCA) methodology. The treatments are either in-situ or ex-situ, the latter requiring an initial dredging to transport the contaminated sediments to the management site. More in detail, four ex-situ remediation technologies based on landfilling, bioremediation, electrokinetic technique and soil washing were identified. These technologies are compared to an in-situ strategy currently under validation for enhancing bioremediation of the polluted sediments of the Bagnoli-Coroglio site. Our results indicate that the disposal in landfilling site is the worst option in most categories (e.g., 650 kg CO2 eq./t of treated sediment, considering the nearest landfilling site), followed by the bioremediation, mainly due to the high energy demand. Electrokinetic remediation, soil washing, and innovative in-situ technology represent the most sustainable options. In particular, the new in-situ technology appears to be the least impacting in all categories (e.g., 54 kg CO2 eq./t of treated sediment), although it is expected to require longer treatment time (estimated up to 12 months based on its potential efficiency). It can reduce the impact on climate change more than 12 times compared to the disposal and 7 times compared to bioremediation in addition to the possibility to avoid/reduce the dredging operations and the consequent dispersion of pollutants. The results open relevant perspectives towards more eco-sustainable and costly effective actions for the reclamation of contaminated marine sediments
Intelligenza artificiale applicata alla radiomica: diagnostica differenziale dei tumori cerebrali intrassiali e risvolti clinici nella loro gestione neurochirurgica
Full text linkI glioblastomi (GBM) rappresentano i tumori cerebrali maligni primari più comuni negli adulti, con un’incidenza stimata di 3–4 casi per 100.000 abitanti per anno e una prognosi sfavorevole nonostante i progressi terapeutici. I linfomi primitivi del sistema nervoso centrale (PCNSL), sebbene più rari, mostrano un’incidenza in aumento e richiedono un approccio terapeutico fondamentalmente diverso, basato principalmente su chemioterapia ad alte dosi e radioterapia. In questo contesto, una diagnosi differenziale accurata e precoce tra GBM e PCNSL è cruciale per garantire strategie terapeutiche personalizzate.
L’obiettivo del nostro studio è stato sviluppare e addestrare un sistema di intelligenza artificiale (AI) basato su analisi radiomica, in grado di differenziare i GBM dai linfomi cerebrali primitivi. Il modello è stato addestrato utilizzando esclusivamente immagini di risonanza magnetica (MRI) pesate in T1 con mezzo di contrasto, estraendo caratteristiche radiomiche avanzate ispirate ai modelli presenti nella letteratura attuale. Il sistema di AI ha raggiunto una sensibilità diagnostica paragonabile, e in alcuni casi superiore, a quella di neuroradiologi esperti, in linea con recenti evidenze sul ruolo dell’AI nell’imaging neuro-oncologico. Questi risultati suggeriscono una potenziale utilità dell’AI come strumento di supporto decisionale in contesti clinici controllati e come ausilio complementare all’interpretazione esperta nella pratica quotidiana.
Nonostante l’utilizzo di una singola sequenza MRI, il modello ha dimostrato prestazioni promettenti, indicando che l’inclusione di ulteriori sequenze multiparametriche potrebbe migliorare ulteriormente l’accuratezza diagnostica negli sviluppi futuri.Glioblastomas (GBM) represent the most common primary malignant brain tumors in adults, with an estimated incidence of 3–4 cases per 100,000 population per year and a poor prognosis despite therapeutic advances. Primary central nervous system lymphomas (PCNSL), although rarer, show an increasing incidence and require a fundamentally different therapeutic approach, primarily based on high-dose chemotherapy and radiotherapy. In this context, an accurate and early differential diagnosis between GBM and PCNSL is crucial to ensure personalized treatment strategies.
The aim of our study was to develop and train an artificial intelligence (AI) system based on radiomic analysis capable of differentiating GBM from primary cerebral lymphomas. The model was trained exclusively using contrast-enhanced T1-weighted magnetic resonance imaging (MRI), extracting advanced radiomic features inspired by current literature models. The AI system achieved a diagnostic sensitivity comparable to, and in some cases exceeding, that of experienced neuroradiologists, consistent with recent reports on the role of AI in neuro-oncologic imaging. These findings suggest the potential utility of AI as a decision-support tool in controlled clinical settings and as a complementary aid to expert interpretation in daily practice.
Despite relying solely on a single MRI sequence, the model demonstrated promising performance, indicating that the inclusion of additional multiparametric MRI sequences could further enhance diagnostic accuracy in future developments
Enhancing hemocompatibility of titanium alloys through plasma immersion ion implantation
No full textTitanium and its alloys have been extensively used in biomedical applications due to their recognized biocompatibility in contact with bone, mechanical strength, and corrosion resistance. However, their use in blood-contacting medical devices, such as coronary stents and valves, is limited, regarding thrombus formation and restenosis. Surface modification has the potential to enhance blood-implant interaction. This study investigates the effects of oxygen plasma immersion ion implantation on the surface properties of pure titanium, Ti6Al4V, and Ti–15Mo alloys, with treatments conducted at two bias voltages (Ubias = −1 kV and −10 kV). Bias voltage was hypothesized to influence surface chemistry, topography, roughness and physical properties, which are key factors in blood protein adsorption and the coagulation cascade. Physicochemical properties were analyzed using scanning electron microscopy (SEM), atomic force microscopy (AFM), contact angle measurements, high-resolution X-ray photoelectron spectroscopy (XPS), and X-ray diffraction (XRD). Hemocompatibility was assessed through hemolysis and clotting time tests. Results demonstrated that blood clotting was significantly delayed within the first 15 min of blood contact for all the treated surfaces, with no hemolytic reactions observed. Among the tested samples, Ti6Al4V modified at a lower bias voltage exhibited the most promising results, forming a smooth, hydrophilic surface with an oxide layer that enhanced hemocompatibility compared to other substrates. These findings reveal the potential of this plasma treatment to optimize the hemocompatibility of titanium alloys, providing a promising pathway for improving the performance of blood-contacting medical devices
Is less more? A meta-analysis comparing single-port and multi-port robotic surgery in children
No full textBackground: Single-port (SP) robotic surgery is a recent advancement in minimally invasive techniques, aiming to reduce surgical trauma compared to multi-port (MP) approaches. Nevertheless, its application in pediatric surgery is not yet well defined. This meta-analysis aims to assess the safety and perioperative outcomes of single-port versus multi-port robotic surgery in the pediatric population. Methods: A systematic review and meta-analysis were performed according to PRISMA guidelines. PubMed, Scopus, Web of Science, Medline, and Cochrane Library were searched up to June 2024 for studies comparing SP and MP robotic surgery in pediatric patients (0–18 years). Primary outcomes included operative time (OT), console time (CT), blood loss (BL), length of hospital stay (LOS), conversion rates, and complications. Results: Five retrospective studies, including 147 pediatric patients (87 MP, 60 SP), were analyzed. No significant differences were found in operative time (WMD = 4.56, 95 % CI [-23.10, 32.22], P = 0.71), console time (WMD = 3.05, 95 % CI [-41.87, 47.96], P = 0.88), or complication rates (OR = 1.32, 95 % CI [0.27, 6.59], P = 0.73). Single-port surgery was linked to lower blood loss (WMD = 26.40, 95 % CI [-102.70, 155.50], P = 0.0005) and a slightly shorter length of hospital stay (WMD = 0.30, 95 % CI [-0.04, 0.55], P = 0.02). Neither group experienced any conversions to open surgery. Conclusions: SP robotic surgery appears to be a feasible and safe alternative to MP approaches in children, with potential benefits in reducing blood loss, mainly demonstrated in pyeloplasty, and hospital stay. However, given the heterogeneity and retrospective nature of the current evidence, larger prospective studies are warranted to confirm these findings
Biotechnological tools for the genetic improvement of cultivated strawberry (Fragaria × ananassa)
No full textLa coltivazione della fragola (Fragaria × ananassa) è sempre più influenzata da fenomeni legati ai cambiamenti climatici, motivo per cui è necessario lo sviluppo di sistemi di propagazione certificati e strategie di miglioramento genetico avanzate. Questo lavoro di dottorato ha studiato soluzioni biotecnologiche basate sulla coltura in vitro e tecnologie del DNA ricombinante perseguendo tre principali obiettivi: i) ottimizzare protocolli di stabilizzazione e rigenerazione in vitro, ii) valutare le performance agronomiche delle piante derivanti da tale approccio a confronto con metodiche tradizionali, e iii) generare linee con minor suscettibilità a patogeni fungini sfruttando l’RNA interferente (RNAi).
L'ipoclorito di sodio è risultato l’agente chimico più efficace per la sterilizzazione e la stabilizzazione di materiale in vitro. La rigenerazione in vitro da foglia più efficiente è stata ottenuta tramite un substrato contenente 0,5 mg/L di thidiazuron e 0,02 mg/L di acido 2,4-diclorofenossiacetico. Il confronto tra piante rigenerate da foglia (RMP), micropropagate (MMP), e piante frigo (FMP) ha evidenziato che le RMP producono più stoloni e cime radicate in sistemi fuori suolo. In pieno campo, la prima generazione delle piante MMP non ha mostrato differenze fenotipiche rispetto alla prima generazione delle FMP. Parallelamente, sono stati ottimizzati protocolli di trasformazione genetica mediata da Agrobacterium tumefaciens per indurre silenziamento genico nell’ospite (HIGS). La trasformazione stabile di costrutti RNAi in fragola, mirati a geni di B. cinerea, C. gloeosporioides e P. aphanis, ha raggiunto un'efficienza fino al 25% nella cultivar Koinè. Le linee selezionate hanno mostrato un'espressione stabile dei trascritti e una significativa riduzione della sintomatologia fungina nei successivi saggi biologici. In conclusione, l'impiego della tecnica HIGS e della coltura in vitro si conferma una strategia promettente per migliorare la sostenibilità della coltivazione della fragola commerciale.Strawberry cultivation (Fragaria × ananassa) is increasingly affected by climate change, requiring the development of certified propagation systems and advanced breeding strategies. This PhD research investigated biotechnological approaches based on in vitro culture and recombinant DNA technologies, focusing on three main objectives: (i) optimization of in vitro stabilization and regeneration protocols; (ii) evaluation of the agronomic performance of plants derived from these approaches compared with traditional propagation methods; and (iii) generation of lines with reduced susceptibility to fungal pathogens through RNA interference (RNAi). Sodium hypochlorite proved to be the most effective chemical agent for the sterilization and stabilization of in vitro plant material. The most efficient leaf-derived in vitro regeneration was achieved using a culture medium supplemented with 0.5 mg·L−1 thidiazuron and 0.02 mg·L−1 2,4-dichlorophenoxyacetic acid. A comparison among leaf-regenerated plants (RMP), micropropagated plants (MMP), and frigo plants (FMP) showed that RMP produced a higher number of runners and rooted tips under soilless cultivation systems. Under open-field conditions, the first generation of MMP did not display phenotypic differences compared with the first generation of FMP. In parallel, Agrobacterium tumefaciens–mediated genetic transformation protocols were optimized to induce host-induced gene silencing (HIGS). Stable transformation of RNAi constructs in strawberry, targeting genes from B. cinerea, C. gloeosporioides, and P. aphanis, achieved up to 25% efficiency in the Koinè cultivar. Selected lines exhibited stable transcript expression and a marked reduction in fungal symptom development in subsequent bioassays. In conclusion, the combined application of HIGS technology and in vitro culture represents a promising strategy to enhance the sustainability of commercial strawberry cultivation
Three essays on Inequality and Climate Change reduction in a Degrowth scenario
No full textQuesto lavoro esplora le potenzialità dell'adozione di un approccio alternativo e sostenibile per affrontare il cambiamento climatico, rispetto alle politiche tradizionali che spesso privilegiano la crescita economica e approcci basati sul mercato. L'alternativa proposta è la decrescita economica sostenibile, intesa come una riduzione volontaria e graduale dei consumi pro capite e delle ore lavorate, accompagnata da politiche di riqualificazione ambientale.
Nel primo capitolo vengono utilizzate tecniche di elaborazione del linguaggio naturale (Natural Language Processing, NLP), per analizzare l'evoluzione del dibattito accademico sulle questioni ambientali, con particolare attenzione al concetto di decrescita.
Il secondo capitolo adotta un modello Quantile Vettoriale Autoregressivo (QVAR) per stimare gli effetti di contagio tra la riduzione delle ore lavorate e dei consumi, disuguaglianze di reddito, l'Impronta Ecologica e l'Impronta Carbonica, con focus sui quantili estremi delle distribuzioni. L'approccio permette di comprendere la risposta dei sistemi economici in scenari di stress ambientale.
Il terzo capitolo propone un indicatore di entropia come proxy per quantificare l'incertezza climatica, valutando l'impatto dei fattori demografici, economici e produttivi su tale incertezza attraverso un modello di regressione su popolazione, ricchezza e tecnologia (Stochastic Impacts by Regression on Population, Affluence, and Technology, STIRPAT).
Complessivamente, la tesi offre una valutazione integrata degli impatti ambientali e sociali di un cambio di paradigma così radicale, fornendo spunti preziosi per i decisori politici per progettare strategie in grado di ridurre il danno ambientale senza compromettere il benessere socio-economico delle popolazioni.This work explores the potential of adopting an alternative, sustainable approach to addressing climate change, as opposed to traditional policies that often prioritize economic growth and a market-based approach. The proposed alternative is sustainable economic degrowth, understood as a voluntary and gradual reduction in per capita consumption and hours worked, accompanied by environmental regeneration policies.
In the first chapter, Natural Language Processing (NLP) techniques are used to analyse the evolution of the academic debate on environmental issues, with a particular focus on the concept of degrowth.
The second chapter employs a Quantile Vector Autoregressive (QVAR) model to estimate the spillover effects between reductions in hours worked and consumption, income inequality, Ecological Footprint, and Carbon Footprint, with a particular focus on extreme quantiles. This approach aims to understand the response of economic systems in scenarios of environmental stress.
The third chapter proposes an entropy indicator as a proxy for climate uncertainty, assessing the impact of demographic, economic, and productive factors on this uncertainty through a Stochastic Impact by Regression on Population, Affluence, and Technology (STIRPAT) model.
Overall, the thesis aims to provide an integrated assessment of the environmental and social impacts of such a radical paradigm shift, offering valuable insights for policymakers who wish to design strategies that can reduce environmental damage without compromising the socio-economic well-being of populations
Heteroresistance in vancomycin-variable Enterococcus faecium ST80 isolate
No full textPurpose: To investigate the gradual genetic changes occurred in heteroresistant vancomycin-variable (VVE) Enterococcus faecium ST80 isolate - belonging to the vanA ST80 nosocomial clone although recovered from marine environment - contributing to its reversion to a resistant phenotype after exposure to increasing concentrations of vancomycin. Methods: WGS of the parental strain and three revertants obtained by exposure to vancomycin was performed. vanA copy number and difference in vanA expression between VVE and revertants were evaluated by RT-qPCR. Heteroresistance was demonstrated by vancomycin E-test and population analysis profiling. Results: The E. faecium JSEG15 isolate, susceptible to vancomycin although carrying a Tn1546-like transposon on 36-kb plasmid, gave rise to three different resistant mutants following gradual passages on increasing vancomycin concentrations. Revertants showed increased vancomycin MICs, with the highest resistance in JSEG15-rev3 (MIC = 128 mg/L). All revertants presented a deleted Tn1546-like on a 34-kb plasmid and JSEG15-rev2 and JSEG15-rev3 also showed an additional chromosomal copy of the vanRS and vanHAX clusters, and higher vanA gene copies and expression. A mutation in the chromosomal D-Ala-D-Ala ligase gene was identified in JSEG15-rev3 which also transferred the vanA carrying plasmid by in vitro conjugation. The parental strain and all revertants exhibited heteroresistance, with higher frequency of resistant subpopulations in the revertants. Conclusion: These findings emphasize the dynamic genetical changes enabling VVE to regain full resistance, underscoring the need for vigilant treatment strategies to address enterococcal infections and for monitoring the spread of clinically relevant strains in the environment for public health protection
The copepod microbiome in the Mediterranean Sea
No full textLe associazioni tra invertebrati marini e microrganismi sono diffuse in tutti gli oceani. Il microbiota può svolgere ruoli cruciali nella digestione, nell’assorbimento dei nutrienti, nella riproduzione, nelle risposte immunitarie e nei meccanismi di difesa della maggior parte degli animali marini, influenzandone così salute e fitness. Nonostante l’importanza ecologica dei copepodi come componenti chiave delle reti trofiche marine e dei cicli biogeochimici, le informazioni sui loro microbiomi restano limitate. È quindi necessario approfondire la comprensione della diversità e della dinamica dei microbiomi dei copepodi. Ciò è particolarmente rilevante nell’attuale contesto di cambiamento globale, in cui la comprensione delle associazioni copepodi-microbi contribuisce a una visione più ampia delle risposte degli ecosistemi marini.
Questa tesi di dottorato si è focalizzata sui copepodi per indagare come habitat, geografia, condizioni ambientali e tratti dell’ospite interagiscano nel modellare i batteri associati. I pattern su larga scala tra habitat, che vanno dai sistemi marini a quelli di acqua dolce, sono stati caratterizzati attraverso una meta-analisi globale di dataset pubblicati. Contemporaneamente, studi nel Mar Mediterraneo hanno rivelato che i microbiomi associati alle specie dominanti di copepodi erano fortemente strutturati da bacini, stagioni, siti di campionamento, risorsa trofica e parametri ambientali. Le analisi dei pellet fecali hanno indicato che le comunità microbiche in essi ospitate, fortemente influenzate da dieta e disponibilità di cibo, possono fungere da proxy del microbioma intestinale dei copepodi e contribuire ai cicli biogeochimici. Infine, i confronti tra copepodi “wild” a diversi stadi di sviluppo, insieme a esperimenti di allevamento, hanno indicato che i microbiomi dei copepodi erano relativamente stabili nel corso del ciclo vitale, supportando l’esistenza di una robusta associazione ospite-microbioma. Segnali di filosimbiosi e la persistenza di specifici taxa del core microbiome, hanno inoltre suggerito che il trasferimento verticale e le associazioni ospite-microbo possano aver contribuito all’adattamento dei copepodi ad ambienti contrastanti. Nel complesso, questi risultati hanno evidenziato i sistemi copepode-microbioma come componenti sensibili degli ecosistemi pelagici e sottolineato l’importanza di comprenderne le risposte a riscaldamento, deossigenazione e inquinamento per prevedere meglio le future traiettorie della produttività dello zooplancton, dei cicli biogeochimici e della salute degli ecosistemi marini.The associations between marine invertebrates and microbes are widespread throughout the oceans. Microbiota can play crucial roles in digestion, nutrient uptake, reproduction, immune responses and defence mechanisms of almost all marine animals, thereby influencing their health and fitness. Despite the ecological importance of copepods as key components of marine food webs and biogeochemical cycles, information on their microbiomes remains limited. A better understanding of the diversity and dynamics of copepod microbiomes, is therefore needed. This is particularly relevant under ongoing global change, where unravelling copepod-microbe associations contributes to a broader comprehension of marine ecosystem responses.
This PhD thesis focused on marine copepods to investigate how habitat, geography, environmental conditions and host traits interact to shape copepod-associated microbiomes. Large-scale patterns across habitats ranging from marine to freshwater systems were characterised through a global meta-analysis of published datasets. Complementary field studies in the Mediterranean Sea revealed that microbiomes of dominant copepod species were strongly structured by basin, season, site, trophic source and environmental parameters. Fecal pellet analyses indicated that the microbial communities they host, which are strongly influenced by diet and food availability, can serve as a proxy for the gut microbiome of copepods and contribute significantly to biogeochemical cycles and carbon export. Finally, comparisons of wild copepods at different developmental stages, together with first-generation rearing experiments, indicated that copepod microbiomes were relatively stable over the life cycle, supporting the existence of a robust host-microbiome association. Signals of phylosymbiosis and the persistence of specific core taxa further suggested that vertical transfer and host–microbe associations may have contributed to copepod adaptation across contrasting environments. Altogether, these findings highlighted copepod–microbiome systems as sensitive components of pelagic ecosystems and underscored the importance of understanding their responses to warming, deoxygenation and pollution in order to better predict future trajectories of zooplankton productivity, biogeochemical cycling and marine ecosystem health