Multiplex network analysis with application to biological high-throughput data

In questa tesi vengono studiate alcune caratteristiche dei network a multiplex; in particolare l'analisi verte sulla quantificazione delle differenze fra i layer del multiplex. Le dissimilarita sono valutate sia osservando le connessioni di singoli nodi in layer diversi, sia stimando le diverse partizioni dei layer. Sono quindi introdotte alcune importanti misure per la caratterizzazione dei multiplex, che vengono poi usate per la costruzione di metodi di community detection . La quantificazione delle differenze tra le partizioni di due layer viene stimata utilizzando una misura di mutua informazione. Viene inoltre approfondito l'uso del test dell'ipergeometrica per la determinazione di nodi sovra-rappresentati in un layer, mostrando l'efficacia del test in funzione della similarita dei layer. Questi metodi per la caratterizzazione delle proprieta dei network a multiplex vengono applicati a dati biologici reali. I dati utilizzati sono stati raccolti dallo studio DILGOM con l'obiettivo di determinare le implicazioni genetiche, trascrittomiche e metaboliche dell'obesita e della sindrome metabolica. Questi dati sono utilizzati dal progetto Mimomics per la determinazione di relazioni fra diverse omiche. Nella tesi sono analizzati i dati metabolici utilizzando un approccio a multiplex network per verificare la presenza di differenze fra le relazioni di composti sanguigni di persone obese e normopeso.

Misura UHF (Ultra High Frequency) delle scariche parziali nei sistemi isolati a gas: metodologie per la verifica della sensibilità di misura

Le apparecchiature di manovra tramite isolamento a gas (GIS) sono in funzione da più di 45 anni e hanno dimostrato un alto livello di affidabilità, con tassi di difetto molto bassi. Tuttavia, il riscontro pratico indica che alcuni dei guasti che si verificano durante il servizio sono legati a difetti del sistema di isolamento. Molti di questi difetti possono essere rilevati tramite diagnostica con scariche parziali (PD). Per la rilevazione PD, viene utilizzato il metodo UHF perché meno sensibile ai disturbi e quindi più facile da gestire in confronto al metodo convenzionale secondo IEC 60270. Un rapporto di Electra pubblicato nel 1999 dal CIGRE Task Force 15/33.03.05 descrive la procedura in due fasi per la verifica della sensibilità del sistema UHF in modo molto generale. Dopo 15 anni dalla sua applicazione, è diventata necessaria una descrizione più dettagliata, sia su misure necessarie al test di laboratorio (Passaggio 1) che sul test in loco (Passaggio 2).

Analisi evolutiva delle radiosorgenti tramite indagine su “compact steep spectrum" (CSS) e “high frequency peakers” (HFP)

in questo lavoro di tesi verrà fornita dapprima una descrizione degli oggetti che andremo a studiare, e cioè i nuclei galattici attivi (AGN), con un richiamo alle strutture più importanti di cui si costituiscono i loro sottogruppi: Quasar e radiosorgenti. Saranno forniti i segni distintivi con cui poter riconoscere queste forme di sorgenti radio e verrà fornito un modello che consenta di concepire il ciclo vitale radiativo a cui sono destinati tali classi di radiogalassie. Si vedrà che il tipo di radiazione emessa è prevalentemente non termico, anch’essa vista come una loro caratteristica peculiare. Si darà quindi una descrizione di questo meccanismo di radiazione ,che corrisponde all’emissione di sincrotrone, e si analizzeranno, facendo riferimento a questa base teorica, dati provenienti da una particolare tipologia di radiosorgenti: le “High Frequency Peakers” (HFP) che, come si evince dal nome stesso, rappresentano delle sorgenti con uno spettro di radiazione con picco riscontrabile alle alte frequenze. Si studieranno i dati raccolti dal catalogo TGSS a 150MHz con la speranza di avere riscontri tali da poter confermare o rigettare il modello evolutivo che prendiamo in considerazione per descrivere il ciclo vitale di questa classe di oggetti astronomici e di cui daremo spiegazione nelle prossime pagine.

Implementation of network entropy algorithms on hpc machines, with application to high-dimensional experimental data

Network Theory is a prolific and lively field, especially when it approaches Biology. New concepts from this theory find application in areas where extensive datasets are already available for analysis, without the need to invest money to collect them. The only tools that are necessary to accomplish an analysis are easily accessible: a computing machine and a good algorithm. As these two tools progress, thanks to technology advancement and human efforts, wider and wider datasets can be analysed. The aim of this paper is twofold. Firstly, to provide an overview of one of these concepts, which originates at the meeting point between Network Theory and Statistical Mechanics: the entropy of a network ensemble. This quantity has been described from different angles in the literature. Our approach tries to be a synthesis of the different points of view. The second part of the work is devoted to presenting a parallel algorithm that can evaluate this quantity over an extensive dataset. Eventually, the algorithm will also be used to analyse high-throughput data coming from biology.

Species identification, distributioin and reproductive interactions of common smoot-hound and blackspotted smooth-hound shark (Genus mustelus) in the Adriatic Sea

Longstanding taxonomic ambiguity and uncertainty exist in the identification of the common (M. mustelus) and blackspotted (M. punctulatus) smooth-hound in the Adriatic Sea. The lack of a clear and accurate method of morphological identification, leading to frequent misidentification, prevents the collation of species-specific landings and survey data for these fishes and hampers the delineation of the distribution ranges and stock boundaries of the species. In this context, adequate species-specific conservation and management strategies can not be applied without risks of population declining and local extinction. In this thesis work I investigated the molecular ecology of the two smooth-hound sharks which are abundant in the demersal trawl surveys carried out in the NC Adriatic Sea to monitor and assess the fishery resources. Ecological and evolutionary relationships were assessed by two molecular tests: a DNA barcoding analysis to improve species identification (and consequently the knowledge of their spatial ecology and taxonomy) and a hybridization assay based on the nuclear codominant marker ITS2 to evaluate reproductive interactions (hybridization or gene introgression). The smooth-hound sharks (N=208) were collected during the MEDITS 2008 and 2010 campaigns along the Italian and Croatian coasts of the Adriatic Sea, in the Sicilian Channel and in the Algerian fisheries. Since the identification based on morphological characters is not strongly reliable, I performed a molecular identification of the specimens producing for each one the cytochrome oxidase subunit 1 (COI) gene sequence (ca. 640 bp long) and compared them with reference sequences from different databases (GenBank and BOLD). From these molecular ID data I inferred the distribution of the two target species in the NC Adriatic Sea. In almost the totality of the MEDITS hauls I found no evidence of species sympatry. The data collected during the MEDITS survey showed an almost different distribution of M. mustelus (confined along the Italian coasts) and M. punctulatus (confined along the Croatian coasts); just one sample (Gulf of Venice, where probably the ranges of the species overlap) was found to have catches of both the species. Despite these data results suggested no interaction occurred between my two target species at least during the summertime (the period in which MEDITS survey is carried out), I still wanted to know if there were inter-species reproductive interactions so I developed a simple molecular genetic method to detect hybridization. This method is based on DNA sequence polymorphism among species in the nuclear ribosomal Internal Transcribed Spacer 2 locus (ITS2). Its application to the 208 specimens collected raised important questions regarding the ecology of this two species in the Adriatic Sea. In fact results showed signs of hybridization and/or gene introgression in two sharks collected during the trawl survey of 2008 and one collected during the 2010 one along the Italian and Croatian coasts. In the case that it will be confirmed the hybrid nature of these individuals, a spatiotemporal overlapping of the mating behaviour and ecology must occur. At the spatial level, the northern part of the Adriatic Sea (an area where the two species occur with high frequency of immature individuals) could likely play the role of a common nursery area for both species.

Seasonality and diel variation in blue whale D calls recorded in the Southern California Bight

Passive acoustic data have been collected using HARPs (High-frequency Acoustic Recording Packages) and were used to assess (1) the seasonality of blue whale D calls in the Southern California Bight, (2) their interannual abundance during 2007-2012 and (3) their diel variation. This goal has been achieved running the GPL (Generalized Power-Law) automated detector. (1) Blue whale D calls were detected in the Southern California Bight from May through November with a peak in July, even though few detections were from December to April as well. A key predictor for blue whale distribution and movement in the California Current region has been identified with zooplankton aggregations, paying a particular attention to those euphausiid species, such as E. pacifica and T. spinifera, which are blue whale favorite krill. The Southern California Bight experiences seasonal upwelling, resulting in an increase of productivity and prey availability. The summer and early fall have been marked as the most favorable periods. This supports the presence of blue whales in the area at that time, supposing these marine mammals exploit the region as a feeding ground. (2) As to the interannual abundance during 2007-2012, I found a large variability. I observed a great increase of vocalizations in 2007 and 2010, whereas a decrease was shown in the other years, which is well marked in 2009. It is my belief that these fluctuations in abundance of D calls detections through the deployed period are due to the alternation of El Nino and La Nina events, which occurred in those years. (3) The assessment of the daily timing of D calls production shows that D calls are more abundant during the day than during the night with a peak at 12:00 and 13:00. Assuming that D calling is associated with feeding, the daily pattern of D calls may be linked to the prey availability. E. pacifica and T. spinifera are among those species of krill which undertake daily vertical migrations, remaining at depth during the day and slowly coming up towards the surface at night. Because of some anatomical arrangements, these euphausiids are very sensitive to the light. Given that we believe D calls have a social function, I hypothesize that blue whales may recognize the hours at the highest solar incidence as the best moment of the day in terms of prey availability, exploiting this time window to advert their conspecifics.

IRAS 00183-7111: ALMA and X-ray view of an ultra luminous infrared galaxy

The aim of this Thesis work is to study the multi-frequency properties of the Ultra Luminous Infrared Galaxy (ULIRG) IRAS 00183-7111 (I00183) at z = 0.327, connecting ALMA sub-mm/mm observations with those at high energies in order to place constraints on the properties of its central power source and verify whether the gas traced by the CO may be responsible for the obscuration observed in X-rays. I00183 was selected from the so-called Spoon diagnostic diagram (Spoon et al. 2007) for mid-infrared spectra of infrared galaxies based on the equivalent width of the 6.2 μm Polycyclic Aromatic Hydrocarbon (PAH) emission feature versus the 9.7 μm silicate strength. Such features are a powerful tool to investigate the contribution of star formation and AGN activity in this class of objects. I00183 was selected from the top-left region of the plot where the most obscured sources, characterized by a strong Si absorption feature, are located. To link the sub-mm/mm to the X-ray properties of I00183, ALMA archival Cycle 0 data in Band 3 (87 GHz) and Band 6 (270 GHz) have been calibrated and analyzed, using CASA software. ALMA Cycle 0 was the Early Science program for which data reprocessing is strongly suggested. The main work of this Thesis consisted in reprocessing raw data to provide an improvement with respect to the available archival products and results, which were obtained using standard procedures. The high-energy data consists of Chandra, XMM-Newton and NuSTAR observations which provide a broad coverage of the spectrum in the energy range 0.5 − 30 keV. Chandra and XMM archival data were used, with an exposure time of 22 and 22.2 ks, respectively; their reduction was carried out using CIAO and SAS software. The 100 ks NuSTAR are still private and the spectra were obtained by courtesy of the PI (K. Iwasawa). A detailed spectral analysis was done using XSPEC software; the spectral shape was reproduced starting from simple phenomenological models, and then more physical models were introduced to account for the complex mechanisms that involve this source. In Chapter 1, an overview of the scientific background is discussed, with a focus on the target, I00183, and the Spoon diagnostic diagram, from which it was originally selected. In Chapter 2, the basic principles of interferometry are briefly introduced, with a description of the calibration theory applied to interferometric observations. In Chapter 3, ALMA and its capabilities, both current and future, are shown, explaining also the complex structure of the ALMA archive. In Chapter 4, the calibration of ALMA data is presented and discussed, showing also the obtained imaging products. In Chapter 5, the analysis and discussion of the main results obtained from ALMA data is presented. In Chapter 6, the X-ray observations, data reduction and spectral analysis are reported, with a brief introduction to the basic principle of X-ray astronomy and the instruments from which the observations were carried out. Finally, the overall work is summarized, with particular emphasis on the main obtained results and the possible future perspectives.

ICE Emissions Estimation through Machine-learning techniques to support real-world R&D experiments

The emissions estimation, both during homologation and standard driving, is one of the new challenges that automotive industries have to face. The new European and American regulation will allow a lower and lower quantity of Carbon Monoxide emission and will require that all the vehicles have to be able to monitor their own pollutants production. Since numerical models are too computationally expensive and approximated, new solutions based on Machine Learning are replacing standard techniques. In this project we considered a real V12 Internal Combustion Engine to propose a novel approach pushing Random Forests to generate meaningful prediction also in extreme cases (extrapolation, very high frequency peaks, noisy instrumentation etc.). The present work proposes also a data preprocessing pipeline for strongly unbalanced datasets and a reinterpretation of the regression problem as a classification problem in a logarithmic quantized domain. Results have been evaluated for two different models representing a pure interpolation scenario (more standard) and an extrapolation scenario, to test the out of bounds robustness of the model. The employed metrics take into account different aspects which can affect the homologation procedure, so the final analysis will focus on combining all the specific performances together to obtain the overall conclusions.

Development of a novel heat treatment for high performances L-PBF Ti6Al4V alloy: microstructural and mechanical characterization

This work presents the experimental development of a novel heat treatment for a high performance Laser Powder Bed Fusion Ti6Al4V alloy. Additive manufacturing production processes for titanium alloys are particularly of interest in cutting-edge engineering fields, however, high frequency laser induced thermal cycles generate a brittle as built microstructure. For this reason, heat treatments compliant with near net shape components are needed before their homologation and usage. The experimental campaign focused on the development of a multi-step heat treatment leading to a bilamellar microstructure. In fact, according to literature, such a microstructure should be promising in terms of mechanical properties both under static and cyclic loads. The heat treatment development has asked for the preliminary analyses of samples annealed and aged in laboratory, implementing several cycles, differing for what concerns temperatures, times and cooling rates. Such a characterization has been carried out through optical and electron microscopy analyses, image analyses, hardness and tensile tests. As a result, the most suitable thermal cycle has been selected and performed using industrial equipment on mini bending fatigue samples with different surface conditions. The same tests have been performed on a batch of traditionally treated samples, to provide with a comparison. This master thesis activity has finally led to the definition of a heat treatment resulting into a bilamellar microstructure, promising in terms of fatigue performances with respect to the traditionally treated alloy ones. The industrial implementation of such a heat treatment will require further improvements, particularly for what concerns the post annealing water quench, in order to prevent any surface alteration potentially responsible for the fatigue performances drop. Further development of the research may also include push-pull fatigue tests, crack grow propagation and residual stresses analyses.