Thesis (Ph. D.)--University of Washington, 2006. / Vita. Includes bibliographical references (leaves 56-75).

Non-canonical base pairs that escape the proof-reading activity of the DNA polymerase emerge from DNA replication as DNA mismatches. To promote genomic integrity, these DNA mismatches are corrected by a secondary protection system, called DNA mismatch repair (MMR). Understanding the details of MMR is important for human health as defects in mismatch repair can result in cancer (e.g. hereditary nonpolyposis colorectal cancer, also known as Lynch syndrome). Being normally stochastic in nature, mismatches can emerge at random locations in a chromosome. Therefore, using a molecular tool to generate substrates for the MMR system at a defined locus has been particularly useful in my study of DNA mismatch repair in vivo. In this study, I have used a CTG•CAG repeat array, also called the “TNR array”, to generate frequent substrates for the MMR system in Escherichia coli. In E. coli, the MMR system searches for hemimethylated GATC motifs around a mismatch to initiate removal of the faulty nascent (un-methylated) strand. Analysing the usage of GATC motifs around the TNR array, I have found that the MMR system preferentially utilizes the GATC motifs on the origin distal side of the TNR array demonstrating that the bidirectionality of MMR in vitro is constrained in live cells. My results suggest that in vivo MMR operates by searching for the nearest hemimethylated GATC site located between the mismatch and the replication fork and excision of the nascent strand occurs directionally away from the fork towards the mismatch. Previous in vitro studies have established that the excision reaction during MMR terminates at a discrete point about 100 bp beyond a mismatch. However, in vivo recombination at a 275 bp tandem repeat, which has been proposed to be mediated by single stranded DNA generated during the excision reaction, has suggested that the end point of the excision reaction in live cells may extend much further from the mismatch than this. I have used this assay for extended excision to determine the influence of GATC sites on excision tracts. In this study, modification of the GATC motifs on the origin proximal side of the TNR has shown that the excision reaction does not stop at a GATC motif on the origin proximal side of the mismatch. In addition, sequential modifications of GATC motifs on the origin distal side of the TNR array, thereby shifting the start point of the excision reaction to a greater distance, have suggested that the length of an excision tract is a function of the distance it covers from the start point rather than from a mismatch. My observation of directionality with respect to DNA replication in the recognition of GATC sites suggested that MMR and DNA replication might be coupled in some way and that perhaps active (or blocked) MMR might impede the progress of the replication fork. However, no replication intermediates were detected using two-dimensional agarose gel electrophoresis of genomic DNA fragment containing the TNR array upon restriction digestion. I was therefore unable to support the hypothesis that active or blocked MMR led to a slowing down of DNA replication. Given my observation of a decrease in MMR by separating the mismatch from the closest origin distal GATC site, I set out to test whether MMR caused any selection pressure for the genomic distribution of GATC motifs. To do this, I generated artificial model genomes using a Markovian algorithm based on the nucleotide composition and codon usage in E. coli. Strikingly, the comparison of the distribution of GATC motifs in the E. coli genome with those from artificial sequences has shown that GATC motifs are distributed randomly in E. coli genome, except for a small clustering effect which has been detected for short spaced (0-40 basepairs) GATC motifs. The observed distribution of slightly over-represented GATC motifs in the E. coli genome appears to be a function of the total number of GATC motifs and it seems that the DNA mismatch repair system has evolved to utilize the natural distribution of GATC motifs to maintain genomic integrity.

Ce travail porte sur la caractérisation et la modélisation des fluctuations aléatoires des paramètres électriques des transistors MOS avancées. La structure de test utilisée est validée expérimentalement au moyen de la méthode de mesure de Kelvin. Pour comprendre le comportement des fluctuations, un modèle est d’abord proposé pour le régime linéaire. Il permet de modéliser les fluctuations de la tension de seuil des transistors avec implants de poche pour toutes les longueurs de transistor et aussi pour toute la gamme de tension de grille. Ensuite, l’appariement du courant de drain est caractérisé et modélisé en fonction de la tension de drain. Pour modéliser les caractéristiques réelles de transistors sans implants de poche, il est nécessaire de considérer la corrélation des fluctuations de la tension de seuil et celles de la mobilité. De plus, des caractérisations sur des transistors avec implants de poche montrent un nouveau comportement de l’appariement du courant de drain. Des caractérisations ont aussi été menées pour analyser l’impact des fluctuations de la rugosité de grille. / This research characterizes and models the mismatch of electrical parameters in advanced MOS transistors. All characterizations are made through a test structure, which is experimentally validated using a structure based on Kelvin method. A model, valid in the linear region, is proposed. It is used for modeling the threshold voltage fluctuations of the transistors with pocket-implants, for any transistor length and gate voltage. It gives a deep understanding of the mismatch, especially for devices with non-uniform channel. Another study analyzes the mismatch of the drain current by characterizing and modeling in terms of the drain voltage. A second model is then proposed for transistors without pocket-implants. In order to apply this model, the correlation of threshold voltage fluctuations and mobility fluctuations must be considered. Characterizations are also performed on transistors with pocket-implants, showing a new drain current mismatch behavior for long transistors. Finally, characterizations are made to analyze the impact of gate roughness fluctuations on mismatch.

Hearing in noise is facilitated by the auditory system’s ability to separate sound into small auditory segments. Separation of sound is achieved using an auditory mechanism called temporal resolution that codes for small silent gaps in an acoustic stimulus. This thesis proposes a new method for measuring temporal resolution and applied it to a small pilot group of individuals with tinnitus. Previous studies have postulated that tinnitus can “fill” in silent gaps thereby making gap detection more difficult. This was shown in studies using the gap prepulse inhibition acoustic startle where the amplitude of a startle response indicates the subject’s ability to detect a small silent gap. However studies using behavioural gap detection do not show significant differences in people with reported tinnitus. Thus the behavioural evidence does not appear to support the hypothesis that tinnitus can “fill” in silent gaps. In this thesis a new method was proposed for measuring neural gap detection: the mismatch negativity response (MMN). The mismatch negativity responses were compared to behavioural measures of gap detection in thirty-five normal hearing adults: five with reported tinnitus and thirty without tinnitus. They underwent recordings to gapped stimuli ranging from 2- to 40-ms gap durations. The stimuli were either a broadband or narrowband noise presented in the absence or presence of a filler noise. Results of these experiments found the broadband and narrowband noises elicited MMNs to silent gaps. The amplitude of the MMN increased with larger gap durations. When filled, the amplitude of the entire waveform was proportionally reduced for all gap durations. However, for the tinnitus group the filler reduced the largest gap durations elicited MMNs amplitudes disproportionately more than for the smaller gap durations. The high and low filler noise reduced the amplitude of the 40-ms gap MMNs. This was not reflective in the behavioural performance of gap detection as there were no significant group differences. These studies show that neural gap detection can be measured using mismatch negativities. Reduced behavioural gap detection performance is reflected by a smaller amplitude of the MMN for suprathrehold gaps. This was shown in both normal hearing participants with elevated behavioural gap detection thresholds and participants with tinnitus. Therefore, electrophysiological recordings to gaps may provide further information on the underlying mechanisms involved in impaired gap detection that may not be captured by behavioural measures alone.

In modern technology exciting developments are related to the ability to understand and control interfaces. Particularly, magnetic interfaces revealing spindependent electron transport are of great interest for modern spintronic devices, such as random access memories and logic devices. From the technological point of view, spintronic devices based on magnetic interfaces enable manipulation of the magnetism via an electric field. Such ability is a result of the different quantum effects arising from the magnetic interfaces (for example, spin transfer torque or spin-orbit torque) and it can reduce the energy consumption as compared to the traditional semiconductor electronic devices. Despite many appealing characteristics of these materials, fundamental understanding of their microscopic properties and related phenomena needs to be established by thorough investigation. In this work we implement first principles calculations in order to study the structural, electric, and magnetic properties as well as related phenomena of two types of interfaces with large potential in spintronic applications: 1) interfaces between antiferromagnetic 3d-metal-oxides and ferromagnetic 3d-metals and 2) interfaces between non-magnetic 5d(4d)- and ferromagnetic 3d-metals. A major difficulty in studying such interfaces theoretically is the typically large lattice mismatch. By employing supercells with Moir e patterns, we eliminate the artificial strain that leads to doubtful results and are able to describe the dependence of the atomic density at the interfaces on the component materials and their thicknesses. After establishing understanding about the interface structures, we investigate the electronic and magnetic properties. A Moir e supercell with transition layer is found to reproduce the main experimental findings and thus turns out to be the appropriate model for simulating magnetic misfit interfaces. In addition, we systematically study the magnetic anisotropy and Rashba band splitting at non-magnetic 5d(4d) and ferromagnetic 3d-metal interfaces and their dependences on aspects such as interdiffusion, surface oxidation, thin film thickness and lattice mismatch. We find that changes of structural details strongly alter the electronic states, which in turn influences the magnetic properties and phenomena related to spin-orbit coupling. Since the interfaces studied in this work have complex electronic structures, a computational approach has been developed in order to estimate the strength of the Rashba band splitting below and at the Fermi level. We apply this approach to the interfaces between a Co monolayer and 4d (Tc, Ru, Rh, Pd, and Ag) or 5d (Re, Os, Ir, Pt, and Au) transition metals and find a clear correlation between the overall size of the band splitting and the charge transfer between the d-orbitals at the interface. Furthermore, we show that the spin splitting at the Fermi surface scales with the induced orbital moment weighted by the strength of the spin-orbit coupling.

The 7th Mismatch Negativity Conference presents the state of the art in methods, theory, and application (basic and clinical research) of the MMN (and related error signals of the brain). Moreover, there will be two pre-conference workshops: one on the design of MMN studies and the analysis and interpretation of MMN data, and one on the visual MMN (with 20 presentations). There will be more than 40 presentations on hot topics of MMN grouped into thirteen symposia, and about 130 poster presentations. Keynote lectures by Kimmo Alho, Angela D. Friederici, and Israel Nelken will round off the program by covering topics related to and beyond MMN.

The P300-based brain-computer interface (BCI) is an extension of the oddball paradigm, and can facilitate communication for people with severe neuromuscular disorders. It has been shown that, in addition to the P300, other event-related potential (ERP) components have been shown to contribute to successful operation of the P300 BCI. Incorporating these components into the classification algorithm can improve the classification accuracy and information transfer rate (ITR). In this paper, a single character presentation paradigm was compared to a presentation paradigm that is based on the visual mismatch negativity. The mismatch negativity paradigm showed significantly higher classification accuracy and ITRs than a single character presentation paradigm. In addition, the mismatch paradigm elicited larger N200 and N400 components than the single character paradigm. The components elicited by the presentation method were consistent with what would be expected from a mismatch paradigm and a typical P300 was also observed. The results show that increasing the signal-to-noise ratio by increasing the amplitude of ERP components can significantly improve BCI speed and accuracy. The mismatch presentation paradigm may be considered a viable option to the traditional P300 BCI paradigm.

La problématique traitée dans la thèse consiste à concevoir des convertisseurs très larges bandes pour les applications stations de base. Le choix d'une architecture à entrelacement temporel a été fait et permet d’augmenter la vitesse des convertisseurs tout en ayant un contrôle sur la consommation. Cependant, les canaux constituant cette architecture évoluent différemment à des variations d'environnement. En conséquence, des erreurs d’appariement entre les canaux dégradent les performances du convertisseur parallèle. Les erreurs les plus connues sont : l’offset, le gain, l’erreur de phase des horloges d’échantillonnage qui sont largement traitées dans la littérature et enfin, l'erreur de bande passantes entre les filtres d'entrées des convertisseurs. Les travaux de la thèse ont permis de proposer des solutions d'estimation et de correction de cette erreur de bande passante afin d'améliorer les performance en linéarité du convertisseur. De plus les techniques de calibrage proposées sont validées à l'aide de circuits réels. / Time-interleaved converter (ti-adc) is an efficient way to increase the speed while maintaining a good accuracy. it consists of the parallelization of several channels; each one running at lower speed. The benefit of this approach is to increase the conversion bandwidth without increasing the power. however, mismatches between the channels cause errors at the digital output and degrade the linearity and the resolution of the system. Gain, offset and clock skew errors are widely treated and we have been interested on bandwidth mismatch error which appears at high frequencies. The goal of the thesis is to develop and implement background calibration techniques for bandwidth mismatch error in a high speed ti-adc (up to 500 msps) in order to achieve a 90 db of sfdr for high input frequencies (up to 385mhz) and up to 94 db at low frequencies. An analog correction solution based on randomization was proposed and a new estimation technique based on gain extraction was implemented and validated for wideband signal.

京都大学 / 0048 / 新制・課程博士 / 博士(医学) / 甲第20073号 / 医博第4166号 / 新制||医||1018(附属図書館) / 33189 / 京都大学大学院医学研究科医学専攻 / (主査)教授 高田 穣, 教授 武藤 学, 教授 武田 俊一 / 学位規則第4条第1項該当 / Doctor of Medical Science / Kyoto University / DFAM

No description available.