Magistro darbe analizuojami aritmetinio loginio įtaiso (ALU) modeliai, operacijos ir architektūros. Išanalizavus mokslinę literatūrą, pasirenkama ALU architektūra bei atliekamų operacijų aibė. Realizuojamas dviejų pakopų sudalintos operacijų aibės ALU modelis. Atlikus apibendrinimą atliekami eksperimentai. Eksperimento modeliai modifikuojami – atliekamas ALU operacijų sudalinimas tarp ALU ir valdančios logikos (CU) operacijų poaibių. Nagrinėjami konkrečių ALU operacijų atlikimo pirmos arba antros pakopos modulyje pranašumai ir trūkumai. Sukurti parametrizuoti (bendriniai) ALU modeliai su kintamo duomenų pločio operandais. ALU modeliai aprašyti aparatūros aprašymo kalba SystemC, sumodeliuoti ir susintezuoti. Gauti ALU SystemC modelių sintezės rezultatai parodė operacijų aibės sudalinimo bei skirtumo tarp retai ir dažnai naudojamų operacijų perkėlimo į skirtingas pakopas poveikį pagrindinėms lusto charakteristikoms: plotui, signalo vėlinimui ir energijos suvartojimui. Optimalus operacijų aibės sudalinimas dviejų pakopų architektūroje leidžia pasiekti reikiamos projektuojamo įtaiso charakteristikos pagerėjimą. Taikant šį metodą, galima kritinei (daug aparatūros resursų naudojančiai) ALU operacijai sudaryti mažesnį plotą luste užimantį arba mažiau energijos suvartojantį ALU variantą. / This work studies arithmetic logic unit (ALU) models, operations and architectures. ALU architecture and operation set is chosen based on the analysis of the known scientific publications. Two-stage divided operation set ALU model is implemented and used for the experiments. The experimental ALU models are modified using different variants of partitioning of ALU operation set, when ALU operations are divided between main ALU block and control unit (CU). Pros and cons of ALU operation performance in the first or the second stage are examined. Developed generic ALU models can be instatiated for data operands with variable data width. ALU models are coded in a high-level hardware description language SystemC, simulated and synthesized. The results of ALU SystemC model synthesis showed the effect of the division of the operation set on the main chip characteristics: area, delay and energy consumption and the difference of subdivision of rare or often used operations into different ALU stages. Optimal subdivision of operation set in two-stage ALU architecture allows getting a better performance of the designed device. Using this method the designer can select an instance of ALU that has a smaller area and consumes less energy for critical (using more hardware resources) operation.

Abstract Lysyl hydroxylase (EC 1.14.11.4, procollagen-lysine 2-oxoglutarate 5-dioxygenase, PLOD) catalyses the formation of hydroxylysine in collagens and in the other collagen like proteins. Hydroxylysine participates in the formation of cross-links between collagen molecules and can bind to the carbohydrates, galactose and glucosylgalactose. Patients with the type VIA Ehlers-Danlos syndrome (EDS) have characteristically a deficiency in hydroxylysine of collagen in their skin that is caused by reduced activity of lysyl hydroxylase 1. In this work the mutations were studied in detail in four different Ehlers-Danlos VIA patients. The first patient characterized in this study had a duplication of seven exons in the lysyl hydroxylase gene 1. The mutation was caused by homologous recombination of two identical 44-nucleotide regions of Alu sequences in introns 9 and 16 in the gene. This study also suggests that uniparental isodisomy does not explain the homozygosity of the mutation. The second patient was found to have two mutations in the gene for lysyl hydroxylase 1 in a compound heterozygote state. The study resulted in the discovery of the first deletion mutation in the gene. The deletion was caused by an Alu-Alu recombination that removes about 3 kb from the gene including all the exon 17 sequences. The other mutation causes deletion of exon 16 from the mRNA. Deletion of the penultimate nucleotide of intron 15 destroys the consensus sequence of the intron/exon boundary and thus causes the deletion. The third patient was described to have a nonsense codon in exon 14 of one allele which causes a reduction in the amount of lysyl hydroxylase mRNA and leads to aberrant RNA splicing in the cell. The other allele was concluded to be operationally null. In the last work two novel null mutations were found in the gene for lysyl hydroxylase 1. The first was a one nucleotide deletion in the acceptor splice site of intron 4 and the other an insertion of a C nucleotide in exon 2. The abnormal alleles lead to markedly decreased lysyl hydroxylase mRNA levels. This work revealed many exon deleted splicing variants of lysyl hydroxylase mRNA which were first discovered in affected cells, but traces of similarly spliced mRNA species were also found in the cytoplasm of normal human skin fibroblasts. These data indicate that the splicing machinery of the cell is leaky. In this thesis, several types of stuctural mutations in the DNA were found to be responsible for lysyl hydroxylase deficiency in patients with type VIA variant of EDS. The different mechanisms causing these mutations were also studied in detail.

Power consumption is a critical design issue in embedded processors. As part of our low power processor design project, this thesis work aims to reduce power consumption on two typical processor components: Register File (RF), and Arithmetic and Logic Unit (ALU). Register File is one of the most power hungry components in the processor, consuming about 20% of the processor power. The ALU is the working horse in the processor, responsible for almost all basic computing operations. Although ALU does not consume as high power as the register file, we observe that it can be power intensive in terms of power dissipation per silicon area unit and may result in a thermal hot spot in the processor. Existing approaches to reduce power on the register file and ALU are effective. However, most of them either entail extensive hardware design efforts, or require a significant amount of work on post-compilation software code modification. The approaches proposed in this thesis avoid such problems. We only customize the internal structure of the processor components and keep the components’ interface to other system parts intact, so that the customization to a component is transparent to its external hardware design and no modification/alteration to other hardware components or to the software code is required. This customization strategy is well suitable to our whole low power processor design project and can be applied to any customization of an existing system for a given application. We have applied our customization approaches to a set of benchmarks in a variety of application domains. Our experimental results show that the power savings on register file are in a range from 18.8% to 45.5%, an average of 29.7% register file power can be saved. For the arithmetic and logic unit, the power savings are from 43.5% to 49.6% and the average saving is 46.9% as compared to the original designs. We also combine the customization of both the ALU and the register file. With the customizing of the ALU and the register file simultaneously, the processor power consumption can be reduced from 3.9% to 10.1%; on average, 6.44% processor power can be saved. Most importantly, the power saving achievement is at the cost of neither hardware complexity nor processor performance, and the implementation is extremely straightforward and can be easily incorporated into a processor design environment, such as ASIPMeister (a design tool, to automatically generate a VHDL model for application specificinstruction set processors) used in our research.

Power consumption is a critical design issue in embedded processors. As part of our low power processor design project, this thesis work aims to reduce power consumption on two typical processor components: Register File (RF), and Arithmetic and Logic Unit (ALU). Register File is one of the most power hungry components in the processor, consuming about 20% of the processor power. The ALU is the working horse in the processor, responsible for almost all basic computing operations. Although ALU does not consume as high power as the register file, we observe that it can be power intensive in terms of power dissipation per silicon area unit and may result in a thermal hot spot in the processor. Existing approaches to reduce power on the register file and ALU are effective. However, most of them either entail extensive hardware design efforts, or require a significant amount of work on post-compilation software code modification. The approaches proposed in this thesis avoid such problems. We only customize the internal structure of the processor components and keep the components’ interface to other system parts intact, so that the customization to a component is transparent to its external hardware design and no modification/alteration to other hardware components or to the software code is required. This customization strategy is well suitable to our whole low power processor design project and can be applied to any customization of an existing system for a given application. We have applied our customization approaches to a set of benchmarks in a variety of application domains. Our experimental results show that the power savings on register file are in a range from 18.8% to 45.5%, an average of 29.7% register file power can be saved. For the arithmetic and logic unit, the power savings are from 43.5% to 49.6% and the average saving is 46.9% as compared to the original designs. We also combine the customization of both the ALU and the register file. With the customizing of the ALU and the register file simultaneously, the processor power consumption can be reduced from 3.9% to 10.1%; on average, 6.44% processor power can be saved. Most importantly, the power saving achievement is at the cost of neither hardware complexity nor processor performance, and the implementation is extremely straightforward and can be easily incorporated into a processor design environment, such as ASIPMeister (a design tool, to automatically generate a VHDL model for application specificinstruction set processors) used in our research.

"Thesis approved for the degree of doctor of science (economics) in the University of London." / "Mono texts": p.[73]-143. Microfilm (positive) (Literature of folklore ; reel 510) Bibliography: p.[5].

No description available.

TAF6 est un co-facteur de transcription qui fait partie du complexe multiprotéique, TFIID. TAF6 possède 4 isoformes et deux d'entre eux ont des fonctions opposées. TAF6[alpha] est l'isoforme anti-apoptotique et peut dimériser avec TAF9, tandis que TAF6[delta] est l'isoforme pro-apoptotique et ne peut pas dimériser avec TAF9. TAF6[delta] induit l'apoptose indépendamment de p53. Cette nouvelle voie apoptotique présente un intérêt certain et l'étude de la régulation de l'épissage alternatif de TAF6[delta] est essentielle pour comprendre le mécanisme par lequel TAF6[delta] induit l'apoptose. Nous avons donc développé un minigène de TAF6 qui récapitule le patron d'épissage endogène. Nous avons muté le minigène TAF6 afin d'identifier des éléments sur l'ARN controllants son épissage. Ces éléments ont été retrouvés surtout au niveau de l'exon 2 et de l'intron entre l'exon 2 et l'exon 3. Trois sites activateurs d'épissage ont été déterminés dans l'exon 2. De plus, nous avons démontré la présence d'une structure secondaire importante dans le choix du site 5' d'épissage. Nous avons aussi établi que la distance de 30 nucléotides entre les deux sites 5' d'épissage avait un rôle à jouer dans le choix de ce site d'épissage. Un dernier site activateur a été positionné dans l'intron. D'après ces résultats de mutations et un criblage PCR d'ADNc de cellules transfectées par des siRNAs contre des protéines de liaison à l'ARN, nous avons identifié hnRNP K comme facteur trans potentiel régulant l'épissage de TAF6. Les facteurs trans liants les éléments cis permettent de réguler en partie l'épissage alternatif d'un ARN pré-messager. Des expériences d'immunoprécipitation in vivo nous ont permis d'établir qu'il existe une interaction entre l'ARN de TAF6 et hnRNP K. Le signal extracellulaire qui agit sur les éléments cis et permet de changer l'épissage alternatif de TAF6 est encore inconnu. Nous avons tenté de le déterminer, mais, pour l'instant, nos hypothèses n'ont pas été confirmées par les expériences menées. L'apoptose, si elle est dérégulée, peut mener à des pathologies. L'expression de TAF6 et l'inclusion d'éléments répétitifs Alus dans son ARN messager ont été démontrés comme étant augmentées dans les cellules tumorales du sein. Nous avons aussi commencé l'étude de l'importance de ces éléments répétitifs dans l'augmentation de l'expression de TAF6 dans les cancers. Nous avons identifié une mutation ponctuelle au niveau de la deuxième séquence répétitive Alu qui peut être retrouvée dans TAF6, mais nous ne connaissons pas encore sa fonction dans la surexpression de TAF6 dans les cancers du sein. Bref, ce mémoire présente les différents résultats obtenus dans le cadre d'expériences dont le but était d'identifier les éléments cis, les facteurs trans, le signal déclenchant l'épissage et le rôle des éléments répétitifs Alus présents dans l'ARN de TAF6.

No description available.

The population genetic study was carried out on 188 unrelated individuals from 5 populations of the Sahel. Relationships of nomadic Fulani to sedentary populations of different linguistic backgrounds and geographic origins were inferred from 16 polymorphic Alu insertions. Bayesian clustering approaches could be applied due to biallelic multilocus nature of the data. Fulani were shown to be divergent from neighbouring sedentary populations (Kassena and Mossi) and similar to Somali of East Africa. In context of already published genetic data, these results could be interpreted as Saharan origin of Fulani diaspora that was caused by Sahara drying out around 6 000 BP. After this initial migration of nomads to West Africa, a primarily female gene flow (integration of females) must have influenced the Fulani population. In contrast to Fulani, Songhai have shown a signal of recent admixture in concordance with historical and linguistic assumptions. KEY WORDS Alu insertions, Fulani, population genetics, Sahel

"Thesis approved for the degree of doctor of science (economics)--in the University of London." / "Mono texts": p.[73]-143. Bibliography: p. [5].