10081 research outputs found
Sort by
Looking at the big picture: defining a method to incorporate multiple pressure into fisheries management considerations
No full textIn the Mediterranean Sea the scientific advice aimed to maintain the long-term productivity of fish stocks is achieved by single species-stock assessment. Ecosystem-oriented advice requires knowledge on the relations between species biology and the environment that surrounds and a method to forecast the biological response to future scenarios. Taking as a case study the Common cuttlefish in the Adriatic Sea, we collected knowledge and we implemented a probabilistic Risk Assessment to describe the sources of error associated to the single species stock assessment. We observe that Bayesian Belief Networks can be used to summarize outputs of ecological models and to link them to expert based conceptual models. We gathered the knowledge on the ecosystem and anthropic pressures and their relationship with biological process by the means of literature review, single species stock assessment, machine learning models and bayesian meta analysis. We then implement a semi-quantitative extension of a risk assessment based on a hierarchical composite indicator describing stock assessment considerations and a Bayesian belief network to model population dynamic and environmental/ecosystem considerations. The proposed approach combines Risk Table, model weighing, ecological models results and Bayesian Belief Network to identify which is the most relevant source of uncertainty in the single species stock assessment. The Bayesian Belief Network is used to model management and environmental scenarios
tracking the risk probability that growth performance of common cuttlefish is impaired. Food
web, and to a less extent temperature, can impact the growth performance of cuttlefish. Furter research is needed to explicitly model the biomass dynamic as a function of alternative biological parameters accounting for food web status
Vascular access in patients with peripheral arterial disease undergoing TAVI: the hostile registry
No full textBACKGROUND
The optimal access route in patients with severe peripheral artery disease (PAD) undergoing TAVI remains undetermined.
OBJECTIVE
To compare clinical outcomes with transfemoral access (TFA), transthoracic access (TTA), and non-thoracic alternative access (TAA) in TAVI patients with severe PAD.
METHODS
Patients with PAD and hostile femoral access (TFA impossible, or possible only after percutaneous treatment) undergoing TAVI at 28 international centers were included in this registry. The primary endpoint was the propensity-adjusted risk of 30-day major adverse events (MAE) defined as the composite of all-cause mortality, stroke/transitory ischemic attack (TIA) or main access site-related VARC 3 major vascular complications. Outcomes were also stratified according to the severity of PAD using a novel risk score (Hostile score).
RESULTS
Among the 1,707 patients included in the registry, 518 (30.3%) underwent TAVI with TFA after percutaneous treatment, 642 (37.6%) with TTA, and 547 (32.0%) with TAA (mostly transaxillary). Compared with TTA, both TFA (adjusted HR=0.58, 95%CI 0.45-0.75) and TAA (adjusted HR=0.60, 95%CI 0.47-0.78) were associated with lower 30-day rates of MAE, driven by fewer access site-related complications. Composite risks at 1 year were also lower with TFA and TAA compared with TTA. TFA compared with TAA was associated with lower 1-year risks of stroke/TIA (adjusted HR=0.49, 95%CI 0.24-0.98), a finding confined to patients with low Hostile scores (Pinteraction=0.049).
CONCLUSIONS
Among patients with PAD undergoing TAVI, both TFA and TAA were associated with lower 30-day and 1-year rates of MAE compared with TTA, but 1-year stroke/TIA rates were higher with TAA compared with TFA
Latin fragments in the Estense Archives: considerations about the context of re-use and catalogue
No full textLa tesi che qui si presenta muove dall’osservazione della collezione dei Frammenti conservata presso l’Archivio di Stato di Modena, fornendo un'analisi dettagliata della sua storia archivistica fin dalle origini e dei percorsi di ricerca esplorati fino ad oggi. Un censimento condotto durante la ricerca ha inoltre rivelato circa 50 nuovi frammenti latini riutilizzati nei registri dell'Archivio Estense, che permangono in situ. Un notevole numero di frammenti dell'ASMo proviene dalle legature dei volumi del complesso archivistico estense, il quale, coprendo il periodo 1317-1797, costituisce il più ampio fondo dell'Archivio di Stato di Modena.
Diversamente dalle ricerche precedenti dedicate a frammenti specifici, questa tesi adotta uno sguardo olistico, esplorando le testimonianze frammentarie legate a un medesimo contesto di riutilizzo. Il nucleo centrale si concentra sulla metodologia adoperata per ricostruire le 'provenienze archeologiche', con un focus sul riutilizzo nei registri dell'Archivio Estense.
La tesi include frammenti documentari, spesso trascurati, offrendo nuove prospettive sullo studio del fenomeno del riuso, soprattutto nei domini estensi. Questo approccio sottolinea inoltre l’interdisciplinarietà intrinseca agli studi dedicati ai frammenti di manoscritti, già considerati prevalentemente secondo un’ottica paleografica e codicologica e qui sottoposti anche a un’indagine dal punto di vista dell’archivistica e della diplomatica.
La tesi comprende un catalogo parziale dei frammenti pergamenacei latini provenienti dall'Archivio Estense, costruito adottando il modello di scheda descrittiva del database digitale Fragmentarium. L'analisi comprende identificazione, datazione, localizzazione dei frammenti, cui si aggiungono aspetti materiali e storico-contestuali del riuso, e il catalogo riferisce, in alcuni casi, la ricostruzione di membra disiecta.
Parallelamente, si è avviato un progetto pilota per la digitalizzazione dei frammenti dell'ASMo in collaborazione con il Centro Studi ARCE dell'Università di Bologna. La tesi include una relazione sulle attività svolte, con riflessioni sulle metodologie adottate durante il progetto di digitalizzazione.Cette recherche porte sur les fragments de manuscrits des Archives d'État de Modène, en particulier tirés des reliures des registres de la Maison d'Este, qui constituent le plus grand fonds documentaire y conservé, datant de 1317 à 1797. Bien que diversifiée, cette documentation reste cohérente grâce à son origine institutionnelle unique.
La thèse débute par une analyse approfondie de l'histoire archivistique de la collection, explorant ses origines et suggérant une expansion idéale avec environ 50 nouveaux fragments latins réutilisés dans les registres de l'Archivio Estense, découverts au cours de cette recherche doctorale. Ainsi, elle rende public une part significative de l'héritage médiéval manuscrit en fragments, jusqu'au present largement inédite.
Contrairement aux recherches antérieures, cette thèse adopte une approche holistique, examinant les éléments fragmentaires liés au même contexte de réutilisation, avec un accent sur les témoins réutilisés dans les reliures des registres de l'Archivio Estense.
La thèse inclut des fragments documentaires, souvent négligés, offrant de nouvelles perspectives sur l'étude du phénomène de réutilisation, notamment dans les domaines estensi. Cette approche souligne également l'interdisciplinarité inhérente aux études consacrées aux fragments de manuscrits, traditionnellement considérés principalement du point de vue paléographique et codicologique, et ici soumis à une enquête également du point de vue de l'archivistique et de la diplomatique.
Un catalogue partiel des fragments de parchemin en latin des Archives de la Maison d'Este est présenté, suivant le modèle de Fragmentarium, incluant l'identification du contenu, la datation, l'origine, ainsi que des aspects matériels et historico-contextuels du remploi, permettant la reconstruction de membra disiecta.
Parallèlement, un projet pilote de numérisation des fragments des Archives d'État de Modène a été mené avec le Centre d'études ARCE de l'Université de Bologne. Un rapport sur les activités du projet est inclus dans la thèse, avec des réflexions sur les méthodologies adoptées.This research explores manuscript fragments preserved at the State Archives of Modena, notably those originating from bindings within the House of Este's archival complex, the State Archives' largest collection spanning 1317 to 1797. Despite diverse document types, the archival complex maintains coherence due to its unified institutional origin.
The thesis meticulously delves into the fragment collection's archival history, identifying origins and proposing expansion through a census, revealing around 50 new Latin fragments within Archivio Estense registers. This study pioneers the exploration of a significant segment of medieval manuscript heritage, preserved fragmentarily and largely unpublished.
Diverging from previous studies focusing on individual fragments or specific groups, this thesis adopts a holistic approach, examining fragments linked by a common reuse context. The focal point is the development of a methodology for reconstructing 'archaeological' provenances, emphasizing witnesses repurposed in Archivio Estense register bindings.
The thesis introduces documentary fragments, often overlooked in fragmentology studies, enabling comprehensive research and shedding light on systematic reuse within the Este family's domains. This approach underscores the interdisciplinary nature inherent in studies dedicated to manuscript fragments, traditionally approached from paleographical and codicological perspectives, now examined through archival and diplomatic studies.
A partial catalog of Latin parchment fragments from the Estense Archives follows the Fragmentarium database's descriptive sheet model. Fragment analysis includes content, dating, location, as far as material and contextual aspects about reuse. Recognizing shared origins enables the reconstruction of membra disiecta.
Simultaneously, a pilot project for digitizing fragments from the State Archives of Modena was carried out in collaboration with the ARCE Studies Center at the University of Bologna. A report on the project's activities is included in the thesis, along with reflections on the adopted methodologies
Applied genomics in livestock: genome-wide association studies and population genomics analysis in pig and rabbit breeds
No full textThe domestication and selection processes in pigs and rabbits have resulted in the constitution of multiple breeds with broad phenotypic diversity. Population genomics analysis and Genome-wide association study analysis can be utilized to gain insights into the ancestral origins, genetic diversity, and the presence of lethal mutations across these diverse breeds. In this thesis, we analysed the dataset obtained from three Italian Pig breeds to detect deleterious alleles. We screened the dataset for genetic markers showing homozygous deficiency using two approaches single marker and haplotype-based approach. Moreover, Genome-wide association study analyses were performed to detect genetic markers associated with pigs' reproductive traits. In rabbits, we investigated the application of SNP bead chip for detection signatures of selection in rabbits using different methods. This analysis was implemented for the first time in different fancy and meet rabbit breeds. Multiple approaches were utilized for the detection of the selection of signatures including Fst analysis, ROH analysis, PCAdapt analysis, and haplotype-based analysis. The analysis in pigs was able to identify five putative deleterious SNPs and nine putative deleterious haplotypes in the analysed Italian Pig breeds. The genomic regions of the detected putative deleterious genomic markers harboring loss of function variants such as the Frameshift variant, start lost, and splice donor variant. Those variants are close to important candidate genes such as IGF2BP1, ADGRL4, and HGF. In rabbits, multiple genomic regions were detected to be under selection of signature. These genomic regions harbor candidate genes associated with coat color phenotype (MC1R, TYR, and ASIP), hair structure (LIPH), and body size (HMGA2 and COL2A1). The described results in rabbits and pigs could be used to improve breeding programs by excluding the deleterious genetic markers carriers and incorporating candidate genes for coat color, body size, and meat production in rabbit breeding programs to enhance desired trait
The right to explanation
No full textThis research investigates the use of Artificial Intelligence (AI) systems for profiling and decision-making, and the consequences that it poses to rights and freedoms of individuals.
In particular, the research considers that automated decision-making systems (ADMs) are opaque, can be biased, and their logic is correlation-based. For these reasons, ADMs do not take decisions as human beings do. Against this background, the risks for the rights of individuals combined with the demand for transparency of algorithms have created a debate on the need for a new 'right to explanation'. Assuming that, except in cases provided for by law, a decision made by a human does not entitle to a right to explanation, the question has been raised as to whether – if the decision is made by an algorithm – it is necessary to configure a right to explanation for the decision-subject.
Therefore, the research addresses a right to explanation of automated decision-making, examining the relation between today’s technology and legal concepts of explanation, reasoning, and transparency. In particular, it focuses on the existence and scope of the right to explanation, considering legal and technical issues surrounding the use of ADMs.
The research analyses the use of AI and the problems arising from it from a legal perspective, studying the EU legal framework – especially in the data protection field. In this context, a part of the research is focused on transparency requirements under the GDPR (namely, Articles 13–15, 22, as well as Recital 71). The research aims to outline an interpretative framework of such a right and make recommendations about its development, aiming to provide guidelines for an adequate explanation of automated decisions. Hence, the thesis analyses what an explanation might consist of, and the benefits of explainable AI – examined from legal and technical perspectives
Integrated approach of water quality monitoring and hydraulic parameter analysis based on building information modeling in water distribution networks
No full textMicrobiological communities in building plumbing and water distribution systems (WDS) are water quality indicators with significant implications for public health, system efficiency and infrastructure maintenance. Thus, monitoring water quality and understanding WDS structures are essential for detecting contamination, assessing risks, and ensuring effective system management. However, some gaps exist, particularly in older buildings that often lack accessible floor plans, complicating WDS understanding. Furthermore, challenges in sharing and managing data on building structures and water quality analysis hinder collaboration and decision-making necessary for an adequate Water Safety Plan, increasing contamination risks. Finally, traditional assessments often focus primarily on microbiological aspects, overlooking hydraulic parameters, limiting comprehensive contamination mitigation, highlighting the need for a more holistic approach. An integrated approach is proposed to address the challenges by innovatively improving water quality and WDS management. This approach combines microbiological and hydraulic parameters analysis, supported by Building Information Modeling (BIM), enhancing coordination among professionals across various disciplines involved in water quality evaluation. It underlines the importance of considering water holistically, departing from the conventional fragmented approach to water quality assessment. By simultaneously analyzing both microbial and hydraulic aspects, water quality can be improved. BIM is a key element by streamlining data sharing and management, other than addressing the lack of correct and updated WDS layouts. In this context, BIM serves as a dynamic, shareable, and consultable model designed to contain several useful information for risk assessment. Integrating microbiological analysis with hydraulic parameter evaluation could ensure optimal water quality. This thesis shows how the proposed approach improves data flow and communication. Developing a model that correctly reproduces the WDS and includes the relevant information has proven useful for planning sampling campaigns, while integrating microbiological and hydraulic analysis identifies malfunctions within the WDS. This approach enhances water quality assessment, safeguarding public health, and optimizing infrastructure maintenance
Molecular profiling and metabolic dysregulation in esophageal adenocarcinoma
No full textEsophageal adenocarcinoma (EAC) is a malignancy with rising incidence rates and poor prognosis in Western countries, with a 5-year survival rate of approximately 35% to 45%. The molecular mechanisms underlying EAC carcinogenesis remain only partially understood. Extensive genomic analyses have identified frequent somatic structural rearrangements, copy number alterations, and single nucleotide mutations, revealing a high mutation burden. The heterogeneity of EAC continues to pose unresolved challenges regarding its classification, prevention, early diagnostic programs, and treatment strategies, ultimately hindering the development of targeted therapies for more personalized treatment approaches. We analyzed the genetic profiles of EAC samples from 164 naïve patients who had not received chemo-radiotherapy. Through high-coverage next-generation sequencing technology and a customized gene panel, we confirmed that TP53 is the most frequently mutated gene, with a mutation rate of 67.1%, further revealing a complex mutational spectrum in EAC, consisting mainly of point mutations but also insertions and deletions. Missense mutations in TP53 were associated with poorer cancer-specific survival (Log Rank P = 0.0197). We identified disruptive mutations in the HNF1α gene in 7 cases, which had not previously been associated with this tumor type, and all accompanied by alterations in other genes. Future studies will be needed to further investigate the role of HNF1α as a tumor suppressor gene to fully elucidate its functions in tumor development and progression. The metabolic profiling of esophageal carcinoma cell lines (OE19 and FLO-1) revealed distinct metabolic dependencies. The FLO-1 cell line exhibits a higher metastatic potential and a predominantly glycolytic metabolism, while OE19 preferentially rely on oxidative phosphorylation and exhibit marked oxidative stress. This difference is reflected in the mitochondrial network, which appears more fragmented and depolarized in OE19 cells compared to FLO-1. These findings suggest a correlation between mitochondrial dynamics and cellular metabolism that may play a role in tumor aggressiveness
Enhancing fall risk assessment through real-world digital mobility biomarkers
No full textFalls are a significant health concern for older adults, leading to severe injuries and loss of independence. The WHO ranks falls as the second leading cause of accidental injury deaths worldwide. Mobility assessment is crucial in evaluating fall risk. Traditional methods involve clinical measures that require specialized settings, but they may not accurately reflect daily mobility complexities. This thesis aimed to develop and validate tools for assessing fall risk through real-world digital mobility biomarkers using wearable sensors. First, this work presented an overview of fall risk factors, traditional assessments, and the potential of real-world digital mobility biomarkers. Then, it focused on the potential role of real-world walking, turning, postural transfers, and balance in fall risk. Finally, the implemented pipelines were applied in the automatic segmentation of the instrumented L-Test (iL-Test) in a multicohort study. Overall, this work demonstrated the potential of digital mobility biomarkers in real-world settings to enhance future fall risk assessments. By addressing gaps in traditional laboratory-based assessments, it contributed to ongoing efforts in developing clinically actionable outcomes. Future research should extend the work presented here in real-world digital mobility biomarkers, explore multivariate models, target multimodal approaches, and expand their application in diverse populations
A digital pathology approach to increase sustainability in diagnostics
No full textCancer of Unknown Primary (CUP) remains one of the most enigmatic and challenging entities in oncology, characterized by diagnostic complexity, molecular heterogeneity, and limited therapeutic options. This study leverages advanced digital pathology and genomic profiling techniques to address these challenges, offering innovative strategies for improving diagnosis and treatment. By integrating liquid biopsy-based genetic characterization with state-of-the-art artificial intelligence (AI) algorithms, the research aims to enhance the understanding and clinical management of CUP. A custom target panel was employed to analyse circulating tumour DNA (ctDNA) from liquid biopsy samples, uncovering somatic and germline mutations and calculating tumour mutation burden (TMB) in CUP patients. These findings provide insights into the genetic landscape of CUP, including the identification of potential therapeutic targets and biomarkers for treatment response. In parallel, we applied a CUP site-of-origin predictive approach to CUP diagnostic slides, with the final goal to provide a primary site indication useful for therapy decision. Histopathological data from whole slide images (WSI) were processed using state-of-the-art AI models to infer the tissue of origin. These models demonstrated robust classification accuracy, emphasizing their potential for improving diagnostic workflows while reducing reliance on invasive biopsies and hazardous chemicals. The study also highlights the environmental and operational benefits of digital pathology, including reduced energy consumption, minimized chemical usage, and streamlined data sharing, paving the way for more sustainable diagnostic practices. Despite the promising results, challenges such as the heterogeneity of the data and the need for independent cohorts underline the importance of further research. By combining genomics, AI-driven pathology, and sustainable practices, this research contributes to the evolving paradigm of precision oncology, offering hope for better outcomes in one of the most elusive cancer types
Future sustainable hydrogen mobility: advanced modelling and control strategies of H2ICE powertrains coupled with environmental impact and cost assessment
No full textThe ambitious targets set by the European Green Deal for a sustainable transition toward greener transportation solutions have underlined the need for new propulsion concepts. Among the possible solutions, Hydrogen-fuelled Internal Combustion Engines (H2ICEs) might play a pivotal role, allowing to decarbonize the sector in a smooth way. A sustainable transition couples environmental aspects with economic and social ones, therefore it is fundamental to evaluate the overall emissions of the vehicles, considering both carbon and pollutants lifecycle emissions while also evaluating the total expenses, to identify the best compromise between environmental sustainability and financial feasibility. This thesis is structured in two different parts: the first one shows how artificial neural networks techniques have been exploited to generate a fast running 0-D engine model capable to replicate a 1-D reference one. Starting from this model, an innovative torque-based control strategy for H2ICE-powerered powertrains has been developed. The latter (together with the engine model) has been integrated in a vehicle model and tested on reference homologation cycles the purpose being to assess the performance of the vehicle and the capability of the control structure to deliver the requested torque, while minimizing consumption and emissions without compromising reliability. In the second part, different hydrogen fuelled vehicles and a battery-electric one are compared in the current and future energetic scenarios in all EU-27 countries in terms of total lifecycle carbon dioxide emissions (considering vehicle production and energy generation/distribution), pollutant emissions (considering fuel/energy production, raw material extraction, exhaust and wear-associated emissions) and total cost of ownership with the aim of identifying the best powertrain configuration for each country