<p>The capability of encrypting top secret information remains as a major research problem in the GGH cryptosystem, which depends on various attacking methods. The early approaches to attacking the GGH cryptosystem mainly relied on special properties of the lattice generated by the vectors of the private key. Consequently, those attacks are not appropriate for general cases.</p> <p>This thesis presents a GGH attacking method for general cases. A lattice basis reduction algorithm is applied to the public key to get a better basis, which is used to decrypt the ciphertext. In the proposed approach, we concentrate on three lattice reduction algorithms: the LLL algorithm, the approximate optimally-reduced algorithm, and the optimally-reduced algorithm. We have implemented a package in MATLAB for the GGH cryptosystem and the three algorithms. We experimented with two groups of experiments and obtained promising results for lattices of low dimensions.</p> / Thesis / Master of Science (MSc)

O desenvolvimento de resistência antimicrobiana e a consequente seleção de microrganismos multirresistentes consolidam-se como grandes ameaças à saúde global. Neste contexto, a busca por novas drogas antimicrobianas/microbicidas é fundamental e compostos como os peptídeos antimicrobianos (AMPs) tornaram-se alvos atraentes. Os AMPs são compostos químicos de massa molar média e grande diversidade estrutural, produzidos por todos os seres vivos e com capacidade de inibir o crescimento de e/ou matar microrganismos. O AMP Cheferina I (Chef I) foi isolado das raízes de Capsella bursa-pastoris e é resultado da proteólise de uma proteína da família das proteínas ricas em glicina, que em plantas estão relacionadas às funções de defesa e cicatrização. O nosso grupo de pesquisa foi pioneiro no desenvolvimento e estudo de análogos truncados amidados deste AMP atípico rico em glicina (67,9%) e histidina (28,6%), que se mostraram ativos frente às diferentes cepas de Candida e a S. cerevisiae pela internalização/ação celular acompanhada de manutenção da integridade da membrana plasmática; o análogo amidado (Chef Ia) e o análogo marcado com 5(6)-carboxifluoresceína/FAM (FAM-Chef Ia) tiveram as suas atividades antifúngicas potencializadas por íons Zn2+. Este trabalho deu continuidade ao estudo do efeito dos íons metálicos divalentes Zn2+, Cu2+, Ca2+ e Mg2+ nas atividades anticandida/fungistática e candidacida/fungicida a diferentes pHs e forças iônicas, estruturas e localizações intracelulares destes análogos. Os resultados na ausência de íons em pH 5,1 revelaram maior atividade do análogo fluorescente em relação à do não fluorescente. Neste mesmo pH, as atividades anticandida e candidacida de Chef Ia foram influenciadas negativamente pelos íons Ca2+ e Mg2+ (2-4 vezes) enquanto que, na presença de íons Zn2+ as atividades anticandida de ambos os análogos foram aumentadas (Chef Ia: 8-64 vezes; FAM-Chef Ia: 4-32 vezes). Os íons Cu2+ aumentaram a atividade anticandida de Chef Ia (2-4 vezes), mas não a do análogo fluorescente, mas as atividades candidacidas de ambos foram melhoradas (Chef Ia: 2-8 vezes; FAM-Chef Ia: 2 vezes). Em pH 5,1, os íons Zn2+ mantiveram a atividade anticandida de Chef Ia em alta força iônica, mas só FAM-Chef Ia exibiu atividade candidacida. Em pH 7,4 ambos análogos foram inativos em baixa e alta forças iônicas na ausência e presença de Zn2+ ou Cu2+. As maiores porcentagens de folhas-&#946;-antiparalelas e dobras foram observadas no espectro de DC de Chef Ia em pH 7,4, sendo que aqueles registrados em pH 5,1 e 7,4 em presença de íons Zn2 e Cu2+ indicaram a formação de quelatos estruturalmente distintos. Ambos os peptídeos são bioquelantes em potencial, sendo as proporções peptídeo: íon obtidas as seguintes: FAM-Chef Ia = 1:2 para Cu2+, 1:10 para Zn2+; Chef Ia = 1:1 para Cu2+. A análise da internalização celular de FAM-Chef Ia permitiu a suposição de dois mecanismos de internalização (translocação direta e endocitose), sendo que nas células vivas a presença de Zn2+ afetou negativamente a translocação direta (p 0,0343) e potencializou a endocitose (p 0,0002). / The development of antimicrobial resistance and the consequent selection of multiresistant microorganisms have become major threats to global health. In this context, the search for new antimicrobial/microbicidal drugs is crucial and the antimicrobial peptides (AMPs) have been seen as attractive targets. AMPs are chemical compounds of medium molecular mass and high structural diversity produced by all living beings, capable of inhibiting the growth of microorganisms and killing them. The AMP Shepherin I (Shep I) was isolated from the roots of Capsella bursa-pastoris, being a bioactive peptide encrypted in a glycine-rich protein from a family that in plants are strictly related to defense and healing functions. Our research group has pioneered the development and study of amidated truncated analogues of this atypical glycine- (67.9%) and histidine-rich (28.6%) AMP, which has shown activity against different strains of Candida and S. cerevisiae through cellular internalization with maintenance of the plasma membrane integrity. The amide analogue (Chef Ia) and its fluorescent analog labeled with 5 (6) - carboxyfluorescein / FAM (FAM-Chef Ia) had their antifungal activities potentiated by Zn2+ ions, so the present work continued examining the effect of the divalent metallic ions Zn2+, Cu2+, Ca2+ and Mg2+ on the anticandidal/fungistatic and candidacidal/fungicide activities at different pHs and ionic forces, structures and intracellular locations of these analogues. The results in the absence of those ions at pH 5.1 revealed that the fluorescently labelled analog was more potent than the nonfluorescent. At the same pH, Shep Ia anticandidal and candidacidal activities were negatively influenced by Ca2+ and Mg2+ ions (2-4 fold), whereas in the presence of Zn2+ ions the anticandidal activities of both analogues were increased (Shep Ia: 8-64 fold, FAM- Shep Ia: 4-32 fold). Cu2+ ions increased Shep Ia anticandidal activity (2-4 fold) but not that of FAM-Shep Ia, nevertheless, the candidacidal activities of both analogues were increased (Shep Ia: 2-8 fold, FAM-Shep Ia: 2 fold). Also at pH 5.1, the Zn2+ ions helped retaining the anticandidal activity of Shep Ia at high ionic strength, although only FAM-Shep Ia exhibited candidacidal activity. At pH 7.4 both analogues were inactive at low and high ionic strengths in the absence or presence of Zn2+ or Cu2+. The highest percentages of antiparallel &#946;-sheet and turns were observed in Shep Ia CD spectrum at pH 7.4, while those recorded at pH 5.1 and 7.4 in the presence of Zn2+ or Cu2+ ions indicated the formation of structurally different chelates. Both peptides are potential biochelates, with the following peptide:ion ratios: FAM-Shep Ia = 1: 2 for Cu2+, 1:10 for Zn2+; Shep Ia = 1: 1 for Cu2+. The analysis of the cellular internalization of FAM-Chef Ia allowed the assumption of two mechanisms of internalization (direct translocation and endocytosis) and in the living cells the presence of Zn2+ negatively affected the direct translocation (p 0.0343) and potentiated endocytosis (p 0.0002).

Mondialement, le carcinome hépatocellulaire (CHC) et le cancer du poumon (CP) constituent un problème majeur de santé publique. Plusieurs études d'association gène candidat ont montré que les SNPs de gènes candidats à la réparation d'ADN et au métabolisme des monocarbones (MMC) influencent le risque de CHC et CP. Toutefois, aucune étude n'a évalué, de façon exhaustive, l'influence des SNPs de la réparation d'ADN ou du MMC avec le risque de CP ou de CHC. Notre étude vise à identifier - à l'aide de deux SNP array de 384 SNPs - les polymorphismes génétiques de la réparation d'ADN et du MMC qui sont prédictifs du risque de CHC chez des Caucasiens cirrhotiques. Nous avons aussi recherché les déterminants génétiques de la réparation d'ADN associés au risque de CP. Nos résultats ont montré que six SNPs du gène BRIP1 (BRCA1interacting protein C?terminal helicase ; rs4986763, rs4986764, rs1557720, rs4986765, rs2191248, et rs11871785) étaient significativement associés au risque de CHC chez les patients porteurs d'une cirrhose d'étiologie virale selon le modèle génétique additif. Après correction de "False Discovery Rate", BRIP1 rs4986764 et rs1557720 étaient significativement associés au risque de CHC. Deux SNPs du MMC situés sur GGH (rs11545076 et rs11545077) étaient significativement associés au risque de CHC chez les patients porteurs d'une cirrhose d'étiologie non virale. Par ailleurs, seul le polymorphisme POLL rs3730477 était associé à un risque accru de CP dans le modèle génétique récessif. La dernière partie de notre étude était consacrée à l'analyse comparée d'hémogrammes et des dommages d'ADN chez des conducteurs de taxi-moto (CTM) de Cotonou - exposés à l?air pollué par le benzène et les HAPs -, et les témoins non exposés. Nos résultats ont montré une réduction significative du nombre des globules blancs, lymphocytes, neutrophiles et plaquettes, avec une misincorporation accrue d'uracile, de 8 oxodG et la présence d'un adduit majeur d'ADN chez les CTM par rapport aux témoins. En conclusion, nous avons identifié six variants sur BRIP1 et deux variants sur GGH associés au risque de CHC sur une cirrhose d'étiologie virale et non virale, respectivement. De plus, nous avons montré que POLL rs3730477 est un prédicteur significatif du risque de cancer du poumon. Une validation de ces résultats dans des cohortes indépendantes s'avère indispensable / Worldwide, hepatocellular carcinoma (HCC) and lung cancer (LC) represent a major public health problem. Previous studies reported associations between single nucleotide polymorphisms (SNPs) in DNA repair or monocarbon metabolism (MCM) genes and LC or HCC risk. However, influences of these SNPs on LC or HCC risk have not been comprehensively evaluated. Our study aimed to identify potential interesting DNA repair and MCM gene variants associated with HCC risk in cirrhotic Caucasians. To this end, we used the Illumina's GoldenGate® technology and performed a comprehensive investigation of 384 SNPs on 94 DNA repair genes and 384 SNPs on 77 MCM genes. This comprehensive SNP-array fine mapping approach was also used to identify potential interesting DNA repair gene variants associated with susceptibility to LC in Caucasians. Our results showed that six variants on BRIP1 gene (BRCA1 interacting protein C-terminal helicase: rs4986763, rs4986764, rs1557720, rs4986765, rs2191248, and rs11871785) were significantly associated with HCC risk in patients carrying hepatitis virus-associated cirrhosis under an additive genetic model. After false discovery rate (FDR) correction for multiple testing, BRIP1 rs4986764 and rs1557720 displayed statistically significant associations with HCC risk. Two SNPs on GGH gene were associated with HCC risk in patients carrying non viral cirrhosis. In our study, only POLL rs3730477 was associated with an increased LC risk under a recessive genetic model (OR=2.81, 95% CI 1.51?5.24). Lastly, we evaluated hematologic changes and levels of DNA adducts, 8-oxodG, dU, and m5dC in Cotonou's motorbike taxi drivers (MBTD) - exposed to air pollution by benzene and polycyclic aromatic hydrocarbons (PAHs) - compared to unexposed controls. Compared to controls, MBTD displayed a significant decrease in the number of white blood cells, lymphocytes, neutrophils and platelets, with the formation of an unknow DNA adduct, whereas uracil misincorporation and 8-oxodG levels in DNA were significantly increased. In conclusion, we identified six variants on BRIP1 gene and two variants on GGH gene that are associated with susceptibility to HCC. In addition, POLL rs3730477 variant was associated with susceptibility to LC. Replication of these findings in independent cohorts is warranted

Aujourd’hui, le renforcement de la sécurité des systèmes de communications devient une nécessité, par anticipation du développement des ordinateurs quantiques et des nouvelles attaques qui en découleront. Cette thèse explore deux techniques complémentaires permettant d’assurer la confidentialité des données transmises sur des liens sans-fils. Dans la première partie de ce travail, nous nous intéressons au schéma de cryptographie à clé publique basée sur des réseaux de points, qui représente une des techniques les plus prometteuses pour la cryptographie post-quantique. En particulier, nous considérons le cryptosystème Goldreich-Goldwasser-Halevi (GGH), pour lequel nous proposons un nouveau schéma utilisant les GLD. Dans la seconde partie de ce travail, nous étudions la sécurité des canaux de diffusion multi-utilisateur, ayant accès à des mémoires de caches, en présence d'un espion. Nous considérons deux contraintes de sécurité: la contrainte de sécurité individuelle et la contrainte de sécurité jointe. Nous dérivons des bornes supérieure et inférieure pour le compromis sécurisé capacité-mémoire en considérant différentes distributions de cache. Afin d'obtenir la borne inférieure, nous proposons plusieurs schémas de codage combinant codage wiretap, codage basé sur la superposition et codage piggyback. Nous prouvons qu'il est plus avantageux d'allouer la mémoire de cache aux récepteurs les plus faibles. / Today, there is a real need to strengthen the communication security to anticipate the development of quantum computing and the eventual attacks arising from it. This work explores two complementary techniques that provide confidentiality to data transmitted over wireless networks. In the first part, we focus on lattice-based public-key cryptography, which is one of the most promising techniques for the post-quantum cryptography systems. In particular, we focus on the Goldreich-Goldwasser-Halevi (GGH) cryptosystem, for which we propose a new scheme using GLD lattices. In the second part of this work, we study the security of multi-user cache-aided wiretap broadcast channels (BCs) against an external eavesdropper under two secrecy constraints: individual secrecy constraint and joint secrecy constraint. We compute upper and lower bounds on secure capacity-memory tradeoff considering different cache distributions. To obtain the lower bound, we propose different coding schemes that combine wiretap coding, superposition coding and piggyback coding. We prove that allocation of the cache memory to the weaker receivers is the most beneficial cache distribution scenario.