An architecture for adaptive replication-based multimedia streaming in P2P networks

This PhD thesis addresses the issue of scalable media streaming in large-scale networking environments. Multimedia streaming is one of the largest sink of network resources and this trend is still growing as testified by the success of services like Skype, Netflix, Spotify and Popcorn Time (BitTorrent-based). In traditional client-server solutions, when the number of consumers increases, the server becomes the bottleneck. To overcome this problem, the Content-Delivery Network (CDN) model was invented. In CDN model, the server copies the media content to some CDN servers, which are located in different strategic locations on the network. However, they require heavy infrastructure investment around the world, which is too expensive. Peer-to-peer (P2P) solutions are another way to achieve the same result. These solutions are naturally scalable, since each peer can act as both a receiver and a forwarder. Most of the proposed streaming solutions in P2P networks focus on routing scenarios to achieve scalability. However, these solutions cannot work properly in video-on-demand (VoD) streaming, when resources of the media server are not sufficient. Replication is a solution that can be used in these situations. This thesis specifically provides a family of replication-based media streaming protocols, which are scalable, efficient and reliable in P2P networks. First, it provides SCALESTREAM, a replication-based streaming protocol that adaptively replicates media content in different peers to increase the number of consumers that can be served in parallel. The adaptiveness aspect of this solution relies on the fact that it takes into account different constraints like bandwidth capacity of peers to decide when to add or remove replicas. SCALESTREAM routes media blocks to consumers over a tree topology, assuming a reliable network composed of homogenous peers in terms of bandwidth. Second, this thesis proposes RESTREAM, an extended version of SCALESTREAM that addresses the issues raised by unreliable networks composed of heterogeneous peers. Third, this thesis proposes EAGLEMACAW, a multiple-tree replication streaming protocol in which two distinct trees, named EAGLETREE and MACAWTREE, are built in a decentralized manner on top of an underlying mesh network. These two trees collaborate to serve consumers in an efficient and reliable manner. The EAGLETREE is in charge of improving efficiency, while the MACAWTREE guarantees reliability. Finally, this thesis provides TURBOSTREAM, a hybrid replication-based streaming protocol in which a tree overlay is built on top of a mesh overlay network. Both these overlays cover all peers of the system and collaborate to improve efficiency and low-latency in streaming media to consumers. This protocol is implemented and tested in a real networking environment using PlanetLab Europe testbed composed of peers distributed in different places in Europe.

A template-assembled synthetic protein surface mimetic of the von Willebrand factor A1 domain inhibits botrocetin-induced platelet aggregation.

Platelet adhesion, the initial step of platelet activation, is mediated by the interaction of von Willebrand factor (VWF) with its platelet receptor, the GPIb-IX complex. The binding of VWF to GPIb-IX is induced either by increased shear stress or by exogenous modulators, such as botrocetin. At a molecular level, this interaction takes place between the A1 domain of VWF and the GPIb alpha chain of the GPIb-IX complex. We report here the design and functional characteristics of a VWF template-assembled synthetic protein (TASP), a chimeric four-helix-bundle TASP scaffold mimicking the surface of the A1 domain. Twelve residues located on helices alpha 3 and alpha 4 in the native A1 domain were grafted onto a surface formed by two neighboring helices of the TASP. VWF TASP was found to inhibit specifically botrocetin-induced platelet aggregation and to bind both botrocetin and GPIb alpha. However, in contrast to the native A1 domain, VWF TASP did not bind simultaneously to both ligands. Modeling studies revealed that the relative orientation of the alpha helices in VWF TASP led to a clash of bound botrocetin and GPIb alpha. These results demonstrate that a chimeric four-helix-bundle TASP as a scaffold offers a suitable surface for presenting crucial residues of the VWF A1 domain; the potential of the TASP approach for de novo protein design and mimicry is thereby illustrated.

p53-induced protein with a death domain (PIDD) isoforms differentially activate nuclear factor-kappaB and caspase-2 in response to genotoxic stress

Cells respond to DNA damage in a complex way and the fate of damaged cells depends on the balance between pro- and antiapoptotic signals. This is of crucial importance in cancer as genotoxic stress is implied both in oncogenesis and in classical tumor therapies. p53-induced protein with a death domain (PIDD), initially described as a p53-inducible gene, is one of the molecular switches able to activate a survival or apoptotic program. Two isoforms of PIDD, PIDD (isoform 1) and LRDD (isoform 2), have already been reported and we describe here a third isoform. These three isoforms are differentially expressed in tissues and cell lines. Genotoxic stress only affects PIDD isoform 3 mRNA levels, whereas isoforms 1 and 2 mRNA levels remain unchanged. All isoforms are capable of activating nuclear factor-kappaB in response to genotoxic stress, but only isoform 1 interacts with RIP-associated ICH-1/CED-3 homologous protein with a death domain and activates caspase-2. Isoform 2 counteracts the pro-apoptotic function of isoform 1, whereas isoform 3 enhances it. Thus, the differential splicing of PIDD mRNA leads to the formation of at least three proteins with antagonizing/agonizing functions, thereby regulating cell fate in response to DNA damage

The CbrA-CbrB two-component regulatory system controls the utilization of multiple carbon and nitrogen sources in Pseudomonas aeruginosa.

A novel two-component system, CbrA-CbrB, was discovered in Pseudomonas aeruginosa; cbrA and cbrB mutants of strain PAO were found to be unable to use several amino acids (such as arginine, histidine and proline), polyamines and agmatine as sole carbon and nitrogen sources. These mutants were also unable to use, or used poorly, many other carbon sources, including mannitol, glucose, pyruvate and citrate. A 7 kb EcoRI fragment carrying the cbrA and cbrB genes was cloned and sequenced. The cbrA and cbrB genes encode a sensor/histidine kinase (Mr 108 379, 983 residues) and a cognate response regulator (Mr 52 254, 478 residues) respectively. The amino-terminal half (490 residues) of CbrA appears to be a sensor membrane domain, as predicted by 12 possible transmembrane helices, whereas the carboxy-terminal part shares homology with the histidine kinases of the NtrB family. The CbrB response regulator shows similarity to the NtrC family members. Complementation and primer extension experiments indicated that cbrA and cbrB are transcribed from separate promoters. In cbrA or cbrB mutants, as well as in the allelic argR9901 and argR9902 mutants, the aot-argR operon was not induced by arginine, indicating an essential role for this two-component system in the expression of the ArgR-dependent catabolic pathways, including the aruCFGDB operon specifying the major aerobic arginine catabolic pathway. The histidine catabolic enzyme histidase was not expressed in cbrAB mutants, even in the presence of histidine. In contrast, proline dehydrogenase, responsible for proline utilization (Pru), was expressed in a cbrB mutant at a level comparable with that of the wild-type strain. When succinate or other C4-dicarboxylates were added to proline medium at 1 mM, the cbrB mutant was restored to a Pru+ phenotype. Such a succinate-dependent Pru+ property was almost abolished by 20 mM ammonia. In conclusion, the CbrA-CbrB system controls the expression of several catabolic pathways and, perhaps together with the NtrB-NtrC system, appears to ensure the intracellular carbon: nitrogen balance in P. aeruginosa.

Regulatory roles of the GacS/GacA two-component system in plant-associated and other gram-negative bacteria.

The sensor kinase GacS and the response regulator GacA are members of a two-component system that is present in a wide variety of gram-negative bacteria and has been studied mainly in enteric bacteria and fluorescent pseudomonads. The GacS/GacA system controls the production of secondary metabolites and extracellular enzymes involved in pathogenicity to plants and animals, biocontrol of soilborne plant diseases, ecological fitness, or tolerance to stress. A current model proposes that GacS senses a still-unknown signal and activates, via a phosphorelay mechanism, the GacA transcription regulator, which in turn triggers the expression of target genes. The GacS protein belongs to the unorthodox sensor kinases, characterized by an autophosphorylation, a receiver, and an output domain. The periplasmic loop domain of GacS is poorly conserved in diverse bacteria. Thus, a common signal interacting with this domain would be unexpected. Based on a comparison with the transcriptional regulator NarL, a secondary structure can be predicted for the GacA sensor kinases. Certain genes whose expression is regulated by the GacS/GacA system are regulated in parallel by the small RNA binding protein RsmA (CsrA) at a posttranscriptional level. It is suggested that the GacS/GacA system operates a switch between primary and secondary metabolism, with a major involvement of posttranscriptional control mechanisms.

Domain Adaptation in remote sensing: increasing the portability of land-cover classifiers

Among the types of remote sensing acquisitions, optical images are certainly one of the most widely relied upon data sources for Earth observation. They provide detailed measurements of the electromagnetic radiation reflected or emitted by each pixel in the scene. Through a process termed supervised land-cover classification, this allows to automatically yet accurately distinguish objects at the surface of our planet. In this respect, when producing a land-cover map of the surveyed area, the availability of training examples representative of each thematic class is crucial for the success of the classification procedure. However, in real applications, due to several constraints on the sample collection process, labeled pixels are usually scarce. When analyzing an image for which those key samples are unavailable, a viable solution consists in resorting to the ground truth data of other previously acquired images. This option is attractive but several factors such as atmospheric, ground and acquisition conditions can cause radiometric differences between the images, hindering therefore the transfer of knowledge from one image to another. The goal of this Thesis is to supply remote sensing image analysts with suitable processing techniques to ensure a robust portability of the classification models across different images. The ultimate purpose is to map the land-cover classes over large spatial and temporal extents with minimal ground information. To overcome, or simply quantify, the observed shifts in the statistical distribution of the spectra of the materials, we study four approaches issued from the field of machine learning. First, we propose a strategy to intelligently sample the image of interest to collect the labels only in correspondence of the most useful pixels. This iterative routine is based on a constant evaluation of the pertinence to the new image of the initial training data actually belonging to a different image. Second, an approach to reduce the radiometric differences among the images by projecting the respective pixels in a common new data space is presented. We analyze a kernel-based feature extraction framework suited for such problems, showing that, after this relative normalization, the cross-image generalization abilities of a classifier are highly increased. Third, we test a new data-driven measure of distance between probability distributions to assess the distortions caused by differences in the acquisition geometry affecting series of multi-angle images. Also, we gauge the portability of classification models through the sequences. In both exercises, the efficacy of classic physically- and statistically-based normalization methods is discussed. Finally, we explore a new family of approaches based on sparse representations of the samples to reciprocally convert the data space of two images. The projection function bridging the images allows a synthesis of new pixels with more similar characteristics ultimately facilitating the land-cover mapping across images.

c-Jun N-terminal kinase binding domain-dependent phosphorylation of mitogen-activated protein kinase kinase 4 and mitogen-activated protein kinase kinase 7 and balancing cross-talk between c-Jun N-terminal kinase and extracellular signal-regulated kinase pathways in cortical neurons

The c-Jun N-terminal kinase (JNK) is a mitogen-activated protein kinase (MAPK) activated by stress-signals and involved in many different diseases. Previous results proved the powerful effect of the cell permeable peptide inhibitor d-JNKI1 (d-retro-inverso form of c-Jun N-terminal kinase-inhibitor) against neuronal death in CNS diseases, but the precise features of this neuroprotection remain unclear. We here performed cell-free and in vitro experiments for a deeper characterization of d-JNKI1 features in physiological conditions. This peptide works by preventing JNK interaction with its c-Jun N-terminal kinase-binding domain (JBD) dependent targets. We here focused on the two JNK upstream MAPKKs, mitogen-activated protein kinase kinase 4 (MKK4) and mitogen-activated protein kinase kinase 7 (MKK7), because they contain a JBD homology domain. We proved that d-JNKI1 prevents MKK4 and MKK7 activity in cell-free and in vitro experiments: these MAPKK could be considered not only activators but also substrates of JNK. This means that d-JNKI1 can interrupt downstream but also upstream events along the JNK cascade, highlighting a new remarkable feature of this peptide. We also showed the lack of any direct effect of the peptide on p38, MEK1, and extracellular signal-regulated kinase (ERK) in cell free, while in rat primary cortical neurons JNK inhibition activates the MEK1-ERK-Ets1/c-Fos cascade. JNK inhibition induces a compensatory effect and leads to ERK activation via MEK1, resulting in an activation of the survival pathway-(MEK1/ERK) as a consequence of the death pathway-(JNK) inhibition. This study should hold as an important step to clarify the strong neuroprotective effect of d-JNKI1.

Convergent genetic architecture underlies social organization in ants.

Complex adaptive polymorphisms are common in nature, but what mechanisms maintain the underlying favorable allelic combinations [1-4]? The convergent evolution of polymorphic social organization in two independent ant species provides a great opportunity to investigate how genomes evolved under parallel selection. Here, we demonstrate that a large, nonrecombining "social chromosome" is associated with social organization in the Alpine silver ant, Formica selysi. This social chromosome shares architectural characteristics with that of the fire ant Solenopsis invicta [2], but the two show no detectable similarity in gene content. The discovery of convergence at two levels-the phenotype and the genetic architecture associated with alternative social forms-points at general genetic mechanisms underlying transitions in social organization. More broadly, our findings are consistent with recent theoretical studies suggesting that suppression of recombination plays a key role in facilitating coordinated shifts in coadapted traits [5, 6].

Bi-planar 2D-to-3D registration in Fourier domain for stereoscopic X-ray motion tracking

In this paper we present a new method to track bonemovements in stereoscopic X-ray image series of the kneejoint. The method is based on two different X-ray imagesets: a rotational series of acquisitions of the stillsubject knee that will allow the tomographicreconstruction of the three-dimensional volume (model),and a stereoscopic image series of orthogonal projectionsas the subject performs movements. Tracking the movementsof bones throughout the stereoscopic image series meansto determine, for each frame, the best pose of everymoving element (bone) previously identified in the 3Dreconstructed model. The quality of a pose is reflectedin the similarity between its simulated projections andthe actual radiographs. We use direct Fourierreconstruction to approximate the three-dimensionalvolume of the knee joint. Then, to avoid the expensivecomputation of digitally rendered radiographs (DRR) forpose recovery, we reformulate the tracking problem in theFourier domain. Under the hypothesis of parallel X-raybeams, we use the central-slice-projection theorem toreplace the heavy 2D-to-3D registration of projections inthe signal domain by efficient slice-to-volumeregistration in the Fourier domain. Focusing onrotational movements, the translation-relevant phaseinformation can be discarded and we only consider scalarFourier amplitudes. The core of our motion trackingalgorithm can be implemented as a classical frame-wiseslice-to-volume registration task. Preliminary results onboth synthetic and real images confirm the validity ofour approach.

Radiometric age constraints on mineral growth, metamorphism, and tectonism of the Gummfluh klippe, Brianconnais domain of the Prealpes, Switzerland

A combined Ar-40/Ar-39, K/Ar, Rb/Sr and stable isotope study has been made of white micas from the Gummfluh klippe (Brianconnais domain of the Prealpes), Switzerland. The klippe consists mainly of Mesozoic to early Tertiary carbonate rocks metamorphosed from anchizonal to epizonal conditions. At the base of the klippe is a 10-50 m thick, ductilely deformed marble mylonite containing deformed authigenic quartz segregations. Stable isotope measurements of the coexisting calcite (deltaO-18SMOW=24.5) and quartz (deltaO-18SMOW=28.4) from the mylonite indicate relatively low temperatures (< 300-degreesC) during mylonitization. Analyses of white mica separates of varying size fractions from the mylonitic rocks by K/Ar and Rb/Sr techniques yield ages between 57 and 103 Ma. This variation is correlated with two parameters, the size of the mineral fraction, and the proportion of 2M1 (more phengitic) to 1M (more muscovitic) polytype in the sample. The K/Ar and Rb/Sr ages are generally younger in the smaller size fractions, which also containless 2M1 phengite. High precision Ar-40/Ar-39 age spectra from different size fractions of these micas record three distinct components, a small Hercynian component (ca. 200-300 Ma), a significant Eoalpine component (64-80 Ma) forming Ar-40/Ar-39 age plateaus, and a very minor Tertiary component (ca. 20-40 Ma). Characterization of the samples by SEM indicates the presence of two white mica populations, a coarser grained, deformed, detrital mica that probably corresponds to the 2M1 phengite and a finer grained neoformed 1M mica. Collectively these observations suggest that the Gummfluh samples contain a mixture of detrital phengites of Hercynian age together with neocrystallized muscovites grown during the late Eoalpine metamorphic event followed by minor argon loss during the Tertiary. The main geologic episode recorded in the Ar-40/Ar-39 age spectra of white micas in the mylonite is of Late Cretaceous/Early Tertiary age (64-80 Ma), representing the first reliable Eoalpine ages ever to be reported from the Prealpes. Contrary to tectonic models, the marble mylonite at the base of the Gummfluh klippe appears to be a Cretaceous thrust plane and not the thrust surface formed during transport of the klippe into its present position from the Penninic Alps during the Tertiary. The late Cretaceous thrust developed during marine sedimentation at a depth of 800 m below the seafloor at temperatures of approximately 280-degrees-C, facilitated by warm fluids along the tectonic discontinuity.

Crystallization and preliminary X-ray diffraction studies of Drosophila melanogaster Gαo-subunit of heterotrimeric G protein in complex with the RGS domain of CG5036.

Regulator of G-protein signalling (RGS) proteins negatively regulate heterotrimeric G-protein signalling through their conserved RGS domains. RGS domains act as GTPase-activating proteins, accelerating the GTP hydrolysis rate of the activated form of Gα-subunits. Although omnipresent in eukaryotes, RGS proteins have not been adequately analysed in non-mammalian organisms. The Drosophila melanogaster Gαo-subunit and the RGS domain of its interacting partner CG5036 have been overproduced and purified; the crystallization of the complex of the two proteins using PEG 4000 as a crystallizing agent and preliminary X-ray crystallographic analysis are reported. Diffraction data were collected to 2.0 Å resolution using a synchrotron-radiation source.

Effects of antiresorptive agents on bone micro-architecture assessed by Trabecular Bone Score in women age 50 and older. The Manitoba Prospective Study

Antiresorptive agents such as bisphosphonates induce a rapid increase of BMD during the 1st year of treatment and a partial maintenance of bone architecture. Trabecular Bone Score (TBS), a new grey-level texture measurement that can be extracted from the DXA image, correlates with 3D parameters of bone micro-architecture. Aim: To evaluate the longitudinal effect of antiresorptive agents on spine BMD and on site-matched spine microarchitecture as assessed by TBS. Methods: From the BMD database for Province of Manitoba, Canada, we selected women age >50 with paired baseline and follow up spine DXA examinations who had not received any prior HRT or other antiresorptive drug.Women were divided in two subgroups: (1) those not receiving any HRT or antiresorptive drug during follow up (=non-users) and (2) those receiving non-HRT antiresorptive drug during follow up (=users) with high adherence (medication possession ratio >75%) from a provincial pharmacy database system. Lumbar spine TBS was derived by the Bone Disease Unit, University of Lausanne, for each spine DXA examination using anonymized files (blinded from clinical parameters and outcomes). Effects of antiresorptive treatment for users and non-users on TBS and BMD at baseline and during mean 3.7 years follow-up were compared. Results were expressed % change per year. Results: 1150 non-users and 534 users met the inclusion criteria. At baseline, users and non-users had a mean age and BMI of [62.2±7.9 vs 66.1±8.0 years] and [26.3±4.7 vs 24.7±4.0 kg/m²] respectively. Antiresorptive drugs received by users were bisphosphonates (86%), raloxifene (10%) and calcitonin (4%). Significant differences in BMD change and TBS change were seen between users and nonusers during follow-up (p<0.0001). Significant decreases in mean BMD and TBS (−0.36± 0.05% per year; −0.31±0.06% per year) were seen for non-users compared with baseline (p<0.001). A significant increase in mean BMD was seen for users compared with baseline (+1.86±0.0% per year, p<0.0018). TBS of users also increased compared with baseline (+0.20±0.08% per year, p<0.001), but more slowly than BMD. Conclusion: We observed a significant increase in spine BMD and a positive maintenance of bone micro-architecture from TBS with antiresorptive treatment, whereas the treatment naïve group lost both density and micro-architecture. TBS seems to be responsive to treatment and could be suitable for monitoring micro-architecture. This article is part of a Special Issue entitled ECTS 2011. Disclosure of interest: M.-A. Krieg: None declared, A. Goertzen: None declared, W. Leslie: None declared, D. Hans Consulting fees from Medimaps.

Molecular cloning of a novel human I-mfa domain-containing protein that differently regulates human T-cell leukemia virus type I and HIV-1 expression.

Regulation of viral genome expression is the result of complex cooperation between viral proteins and host cell factors. We report here the characterization of a novel cellular factor sharing homology with the specific cysteine-rich C-terminal domain of the basic helix-loop-helix repressor protein I-mfa. The synthesis of this new factor, called HIC for Human I-mfa domain-Containing protein, is controlled at the translational level by two different codons, an ATG and an upstream non-ATG translational initiator, allowing the production of two protein isoforms, p32 and p40, respectively. We show that the HIC protein isoforms present different subcellular localizations, p32 being mainly distributed throughout the cytoplasm, whereas p40 is targeted to the nucleolus. Moreover, in trying to understand the function of HIC, we have found that both isoforms stimulate in T-cells the expression of a luciferase reporter gene driven by the human T-cell leukemia virus type I-long terminal repeat in the presence of the viral transactivator Tax. We demonstrate by mutagenesis that the I-mfa-like domain of HIC is involved in this regulation. Finally, we also show that HIC is able to down-regulate the luciferase expression from the human immunodeficiency virus type 1-long terminal repeat induced by the viral transactivator Tat. From these results, we propose that HIC and I-mfa represent two members of a new family of proteins regulating gene expression and characterized by a particular cysteine-rich C-terminal domain.

Characterization of HbpR binding by site-directed mutagenesis of its DNA-binding site and by deletion of the effector domain.

In the presence of 2-hydroxybiphenyl, the enhancer binding protein, HbpR, activates the sigma54-dependent P(hbpC) promoter and controls the initial steps of 2-hydroxybiphenyl degradation in Pseudomonas azelaica. In the activation process, an oligomeric HbpR complex of unknown subunit composition binds to an operator region containing two imperfect palindromic sequences. Here, the HbpR-DNA binding interactions were investigated by site-directed mutagenesis of the operator region and by DNA-binding assays using purified HbpR. Mutations that disrupted the twofold symmetry in the palindromes did not affect the binding affinity of HbpR, but various mutations along a 60 bp region, and also outside the direct palindromic sequences, decreased the binding affinity. Footprints of HbpR on mutant operator fragments showed that a partial loss of binding contacts occurs, suggesting that the binding of one HbpR 'protomer' in the oligomeric complex is impaired whilst leaving the other contacts intact. An HbpR variant, devoid of its N-terminal sensing A-domain, was unable to activate transcription from the hbpC promoter while maintaining protection of the operator DNA in footprints. Wild-type HbpR was unable to activate transcription from the hbpC promoter when delta A-HbpR was expressed in the same cell, suggesting the formation of (repressing) hetero-oligomers. This model implies that HbpR can self-associate on its operator DNA without effector recognition or ATP binding. Furthermore, our findings suggest that the N-terminal sensing domain of HbpR is needed to activate the central ATPase domain rather than to repress a constitutively active C domain, as is the case for the related regulatory protein XylR.

Evolution of the ferric reductase domain (FRD) superfamily: modularity, functional diversification, and signature motifs.

A heme-containing transmembrane ferric reductase domain (FRD) is found in bacterial and eukaryotic protein families, including ferric reductases (FRE), and NADPH oxidases (NOX). The aim of this study was to understand the phylogeny of the FRD superfamily. Bacteria contain FRD proteins consisting only of the ferric reductase domain, such as YedZ and short bFRE proteins. Full length FRE and NOX enzymes are mostly found in eukaryotic cells and all possess a dehydrogenase domain, allowing them to catalyze electron transfer from cytosolic NADPH to extracellular metal ions (FRE) or oxygen (NOX). Metazoa possess YedZ-related STEAP proteins, possibly derived from bacteria through horizontal gene transfer. Phylogenetic analyses suggests that FRE enzymes appeared early in evolution, followed by a transition towards EF-hand containing NOX enzymes (NOX5- and DUOX-like). An ancestral gene of the NOX(1-4) family probably lost the EF-hands and new regulatory mechanisms of increasing complexity evolved in this clade. Two signature motifs were identified: NOX enzymes are distinguished from FRE enzymes through a four amino acid motif spanning from transmembrane domain 3 (TM3) to TM4, and YedZ/STEAP proteins are identified by the replacement of the first canonical heme-spanning histidine by a highly conserved arginine. The FRD superfamily most likely originated in bacteria.