• Refine Query
  • Source
  • Publication year
  • to
  • Language
  • 317
  • 222
  • 69
  • 44
  • 31
  • 14
  • 12
  • 7
  • 6
  • 6
  • 4
  • 3
  • 3
  • 3
  • 2
  • Tagged with
  • 854
  • 136
  • 124
  • 78
  • 66
  • 61
  • 59
  • 57
  • 49
  • 44
  • 43
  • 41
  • 40
  • 39
  • 38
  • About
  • The Global ETD Search service is a free service for researchers to find electronic theses and dissertations. This service is provided by the Networked Digital Library of Theses and Dissertations.
    Our metadata is collected from universities around the world. If you manage a university/consortium/country archive and want to be added, details can be found on the NDLTD website.
201

Temporal sequence effects: a memory framework

Montgomery, Nicole Votolato 22 June 2007 (has links)
No description available.
202

Incremental Packing Problems: Algorithms and Polyhedra

Zhang, Lingyi January 2022 (has links)
In this thesis, we propose and study discrete, multi-period extensions of classical packing problems, a fundamental class of models in combinatorial optimization. Those extensions fall under the general name of incremental packing problems. In such models, we are given an added time component and different capacity constraints for each time. Over time, capacities are weakly increasing as resources increase, allowing more items to be selected. Once an item is selected, it cannot be removed in future times. The goal is to maximize some (possibly also time-dependent) objective function under such packing constraints. In Chapter 2, we study the generalized incremental knapsack problem, a multi-period extension to the classical knapsack problem. We present a policy that reduces the generalized incremental knapsack problem to sequentially solving multiple classical knapsack problems, for which many efficient algorithms are known. We call such an algorithm a single-time algorithm. We prove that this algorithm gives a (0.17 - ⋲)-approximation for the generalized incremental knapsack problem. Moreover, we show that the algorithm is very efficient in practice. On randomly generated instances of the generalized incremental knapsack problem, it returns near optimal solutions and runs much faster compared to Gurobi solving the problem using the standard integer programming formulation. In Chapter 3, we present additional approximation algorithms for the generalized incremental knapsack problem. We first give a polynomial-time (½-⋲)-approximation, improving upon the approximation ratio given in Chapter 2. This result is based on a new reformulation of the generalized incremental knapsack problem as a single-machine sequencing problem, which is addressed by blending dynamic programming techniques and the classical Shmoys-Tardos algorithm for the generalized assignment problem. Using the same sequencing reformulation, combined with further enumeration-based self-reinforcing ideas and new structural properties of nearly-optimal solutions, we give a quasi-polynomial time approximation scheme for the problem, thus ruling out the possibility that the generalized incremental knapsack problem is APX-hard under widely-believed complexity assumptions. In Chapter 4, we first turn our attention to the submodular monotone all-or-nothing incremental knapsack problem (IK-AoN), a special case of the submodular monotone function subject to a knapsack constraint extended to a multi-period setting. We show that each instance of IK-AoN can be reduced to a linear version of the problem. In particular, using a known PTAS for the linear version from literature as a subroutine, this implies that IK-AoN admits a PTAS. Next, we study special cases of the generalized incremental knapsack problem and provide improved approximation schemes for these special cases. In Chapter 5, we give a polynomial-time (¼-⋲)-approximation in expectation for the incremental generalized assignment problem, a multi-period extension of the generalized assignment problem. To develop this result, similar to the reformulation from Chapter 3, we reformulate the incremental generalized assignment problem as a multi-machine sequencing problem. Following the reformulation, we show that the (½-⋲)-approximation for the generalized incremental knapsack problem, combined with further randomized rounding techniques, can be leveraged to give a constant factor approximation in expectation for the incremental generalized assignment problem. In Chapter 6, we turn our attention to the incremental knapsack polytope. First, we extend one direction of Balas's characterization of 0/1-facets of the knapsack polytope to the incremental knapsack polytope. Starting from extended cover inequalities valid for the knapsack polytope, we show how to strengthen them to define facets for the incremental knapsack polytope. In particular, we prove that under the same conditions for which these inequalities define facets for the knapsack polytope, following our strengthening procedure, the resulting inequalities define facets for the incremental knapsack polytope. Then, as there are up to exponentially many such inequalities, we give separation algorithms for this class of inequalities.
203

Reliability diagnostic strategies for series systems under imperfect testing

Reller, Susan R. 20 November 2012 (has links)
An expected cost model was developed for failure detection in series systems under imperfect testing. Type I and type II error probabilities are included and single-pass sample paths are required. The model accounts for the expected costs of testing components, false positive termination, and no-defect-found outcomes. Based on the model, a heuristic was developed to construct the cost minimizing testing sequence. The heuristic algorithm utilizes elementary arithmetic computations and has been successfully applied to a variety of problems. Furthermore, the algorithm appears to be globally convergent. Choice of a starting solution affects the rate of convergence, and guidelines for selecting the starting solution were discussed. Implementation of the heuristic was illustrated by numerical example. / Master of Science
204

The order in which you cope matters: An examination of the moderating role of coping sequence on the impact of stressor type on affect

Minton, Brandon Tyler 08 May 2023 (has links)
To date, few studies have sought to investigate whether the sequence in which individuals engage in coping strategies could impact the effectiveness of those strategies. The present study utilizes an EMA data collection approach to obtain a sample of N = 93 student participants to investigate this potential impact. I analyzed the data with a type of multilevel structural equation model (MSEM) called a cross-lagged panel model (CLPM), where the individual served as the higher level and surveys collected at various time points (three per day for five consecutive days) served as the lower level nested within those individuals. Autoregressive, cross-lagged, and moderation paths were tested to see which constructs at time point T-1 were significantly related to positive affect and negative affect at time point T. Findings indicated more significant relationships for positive affect at time point T than negative affect at time point T. Among these were moderation effects of coping strategy on the relationship between the presence of an interpersonal stressor and positive affect, such that emotion-focused coping buffers that effect and problem-focused coping amplifies it. / Doctor of Philosophy / The history of stress and coping research lends itself well to the consideration of coping with stress as a dynamic process that has effects at later times. Different coping strategies (problem-focused, emotion-focused, and avoidant) may be differentially effective for coping with different types of stressors. This study consists of measuring stressor type, coping strategy, positive affect, and negative affect at 15 different time points, specifically to track the impact of the presence of a stressor, the use of certain coping strategies, and positive and negative affect at any given time point on positive and negative affect at a subsequent time point. Specifically, I hypothesize that stressors decrease positive affect and increase negative affect, and that coping strategies can either buffer or intensify these effects. Support is found for the idea that emotion-focused coping buffers an interpersonal stressor's tendency to decrease positive affect. Support is also found for the idea that problem-focused coping intensifies this same tendency.
205

Matching Genetic Sequences in Distributed Adaptive Computing Systems

Worek, William J. 22 August 2002 (has links)
Distributed adaptive computing systems (ACS) allow developers to design applications using multiple programmable devices. The ACS API, an API created for distributed adaptive com-puting, gives developers the ability to design scalable ACS systems in a cluster networking environment for large applications. One such application, found in the field of bioinformatics, is the DNA sequence alignment problem. This thesis presents a runtime reconfigurable FPGA implementation of the Smith-Waterman similarity comparison algorithm. Additionally, this thesis presents tools designed for the ACS API that assist developers creating applications in a heterogeneous distributed adaptive computing environment. / Master of Science
206

Identification And Characterisation Of Two Silencing Barrier Sequences In Saccharomyces Cerevisiae

Biswas, Moumita 02 1900 (has links)
In eukaryotic cells, genomic DNA exists as chromatin in association with histone octamers called nucleosomes, and various other chromatin proteins. Chromatin structure varies along the chromosome and this influences the state of gene expression. Based on such variations in structure and gene expression, chromatin can be broadly classified into euchromatin (transcriptionally active) and heterochromatin (silent or transcriptionally repressed). In the budding yeast, Saccharomyces cerevisiae, there are four canonical transcriptionally silent regions, namely, the HMR, the HML (cryptic mating loci), the telomeres and the RDN1. Silencing at the HM loci and the telomeres is very well characterized. The repressive structure at the HMR spans around 3.5 Kb and extends between the two silencers E and I. It is well established that silencing in HMR is due to a specialized chromatin organization brought about by Orc1p, Rap1p, Abf1p and Sir proteins. Following recruitment, the Sir proteins spread along the DNA to form a repressive chromatin domain believed to arise from the deacetylation of amino-terminal tails of histones H3 and H4 by Sir2p (an NAD dependent deacetylase) and the interaction of Sir3p and Sir4p with the histones. The bi-directional spreading of silencing at HMR is restricted by barrier or boundary elements that flank the silencers. A tRNAThr gene in the right boundary of HMR acts as a strong barrier. Mutations in the promoter of this tRNA gene (tDNA) or in RNA polymerase III subunits/ transcription factors weaken the barrier activity of this tDNA. The barrier activity of this tDNA is also dependent on histone acetyltransferases like Sas2p and Gcn5p. Silencing in HML is uniformly high between the silencers E and I and falls sharply outside I. Recently, barriers to HML silencing have been discovered. A 0.71Kb sequence near E, which maps to the upstream activating sequence of YCL069W, acts as a robust barrier to spread of HML silencing. This is effectively the left boundary of silent HML. The right boundary maps to the promoter of CHA1 gene though silencing is believed to terminate at HML-I. An unusual form of silencing occurs at the RDN1, which contains 100-150 copies of tandemly repeated rRNA genes. Some RNA polymerase II transcribed genes integrated within the array are silenced by a Sir2p dependent mechanism whereas genes driven by RNA polymerases III and I are transcriptionally active. Though all the three forms of silencing (RDN1, HM and telomere) require Sir2p, RDN1 silencing differs from the others in its relative strength and factors responsible for repression. Several trans-acting factors required for RDN1 silencing are known. However, it is still unclear as to what limits the spread of RDN1 heterochromatin into neighbouring essential genes. RDN1 silencing spreads unidirectionally in its left hand side sequence. However, the zone of RDN1 heterochromatin does not engulf the essential gene, ACS2, which is present ~3 kb away from NTS1. This implies that there is a mechanism by which rDNA heterochromatin is contained. There could be several ways by which this is accomplished. Firstly, the cell could be critically maintaining the levels of Sir2p, the protein required for silencing at all the four silenced loci, such that silencing in the left flank of RDN1 does not spread beyond 300 bp of NTS1 (Buck et al, 2002). There is a ~2.5 kb gene free intervening sequence between NTS1 of the rDNA array and the Ty1 LTR, in which interval Sir2p level could fall below the threshold mark required for causing repression. In fact Buck et al. have demonstrated that Sir2p is bound to upto 1.5 kb from the NTS1 in the left flank but there is no accompanying silencing of the mURA3 reporter in these regions (1200L and 2000L), suggesting that the level of Sir2p at these sequences could be lower than the threshold required for initiation of silencing. Secondly, there could be cis-acting boundary elements or barriers as in the case of HMR, which prevents the propagation of RDN1 silencing. The third option is that termination of RNA polymerase I transcription at the terminator sites automatically halts the spread of rDNA silencing since Buck et al. have demonstrated that progression of rDNA heterochromatin is dependent on RNA pol I transcription. This however, does not seem to be the case as deletion of both the terminator sites within NTS1 does not lengthen the zone of silencing. Finally, there could be an euchromatin organizing center further from the array, which creates an “open” chromatin configuration required to confront the Sir2p mediated condensed chromatin. The balance of these two opposing activities, much like that at the telomeres, could set up a molecular boundary for containing rDNA silent chromatin. We have attempted to identify whether there are any sequences in the unique left flank of RDN1 that can act as a heterochromatin barrier. Towards that end we tested four overlapping fragments from NTS1 of RDN1 to the promoter of ACS2 for boundary activity in a quantitative mating assay. We have found that of all the four fragments tested, only a 0.427 kb tRNAGln-Ty1 LTR fragment, which is present 2.4 Kb from the NTS1 acts as a robust barrier in this assay. Further mapping revealed that the barrier activity of this sequence resides in the tRNAGln gene and that its activity is orientation-independent. tDNAs are transcribed by RNA polymerase III from internal promoters termed Box A and Box B. It has been shown for the HMR-tRNAThr that the transcriptional potential of the tDNA is crucial for its barrier function. Mutations in genes encoding various subunits of the RNA polymerase III complex, or transitions in the conserved bases within Box B known to disrupt transcription complex assembly and subsequent transcription, abrogate the barrier activity of HMR-tRNAThr. Similarly, loss of transcriptional ability of the tRNAAla in the centromere of S. pombe also abolishes its barrier activity, enforcing the fact that RNA polymerase III transcription is a decisive factor for a tDNA barrier. Contrary to the above observations, we report that barrier activity of tRNAGln is very negligibly dependent on RNA polymerase III mediated transcription. Mating assays done with the RNA pol III mutants and promoter point mutants, G18C and C55G in boxes A and B respectively, underline the fact that for this tDNA barrier, RNA pol III driven transcription is dispensable. We also show by RT-PCR analysis that in the C55G tRNAGln mutant there is loss of transcription as expected, whereas other wild type copies of tRNAGln are transcribed. Studies with another tDNA barrier, TRT2-tRNAThr, yielded similar results, again emphasizing the point that transcription through the tDNA, which leads to nucleosome displacement and therefore barrier activity, may not be applicable for all tDNA barriers. Acetylation of amino terminal tails of histones is known to influence the epigenetic state of chromatin. Addition of acetyl moiety to histones H3 and H4 initiates a cascade of events, which involves recruitment of a host of other chromatin modifiers to the target sequence, and ultimately culminates in the formation of an euchromatin-favouring environment. As reported for the HMR right boundary, we find that barrier activity of tRNAGln depends on two histone acetyl transferase complexes, SAS-I (comprised of Sas2p, Sas4p and Sas5p) and SAGA (contains Gcn5p HAT). Contrary to the HMR boundary, the barrier activity of tRNAGln is independent of two other nucleoplasmic HATs, NuA3 (Sas3p being the HAT) and NuA4 (Esa1p is the HAT). Barrier function of TRT2-tRNAThr also depends on HATs. Therefore it appears that requirement of HATs for boundary activity is a conserved theme, albeit with differential effects at different barrier sequences. We next attempted to determine the function of tRNAGln in its natural location on chromosome XII. As mentioned earlier, RDN1 silencing spreads upto ~0.3 kb in its left flanking sequence. However, Sir2p occupancy has been observed till 1.5 kb although there is no silencing of reporter genes observed beyond 0.3 kb of NTS1. This lead us to speculate that there could be a boundary sequence in the left flank that stops silencing, or a euchromatin-organizing element, which counters the propagation of silencing by a long-range effect. Since over expression of Sir2p extends the domain of silencing from 0.3 kb to 2.0 kb and the tRNAGln is present at 2.3 kb from NTS1, it was a good candidate for a heterochromatin barrier/ euchromatiniser. However, deletion of tRNAGln does not affect the zone of RDN1 silencing as is evident from our cell viability assays (which is a measure of the expression of the essential gene ACS2 situated further to the left of tRNAGln). Deletion of SAS2 and GCN5, factors that are required for barrier activity of tRNAGln in mating assays, also have no effect on the extent of spreading of RDN1 silencing in normal or Sir2p over expression conditions. Together, these observations imply that in situ, tRNAGln does not act as a barrier or an element with long-range euchromatin inducing properties. It still remains unclear as to what contains RDN1 silencing. It is possible that the cell critically monitors the level of Sir2p in order to maintain boundaries of silencing at the rDNA locus. Telomeres also nucleate the formation of silenced domain which spreads along the subtelomeric region upto ~ 2Kb. The key players in the formation of telomeric heterochromation are the Sir proteins, Sir2p, Sir3p and Sir4p, Rap1p, yKu complex and ORC. Protein-protein interactions between the telosome and the subtelomeric repeat bound silencing proteins create a domain of core heterochromatin that spreads in the adjacent sequences. While Sir2p deacetylates H4K16, Sir3p interacts with the hypoacetylated histone tails and helps in the spreading of the repressive chromatin structure. As a result telomere proximal genes are silent whereas the ones further away are expressed. There is a gradient of acetylation of histone H4, with the hypoacetylated histones at the telomeric ends and the hyperacetylated ones distant from the telomere. Recently it has been shown that this gradient is maintained by the concerted and antagonistic actions of Sir2p and Sas2p. In a sas2Δ strain Sir3p spreads to ~15 kb in the subtelomeric regions and there is increase in the levels of hypoacetylated histones. Though the molecular mechanism by which telomeric silencing is restrained is beginning to be understood, it remains unanswered whether there are any cis-acting sequences, capable of recruiting euchromatin-inducing factors such as Sas2p, near the telomeres. We have identified a RNA polymerase II driven gene, AAD3, in the subtelomeric region of chromosome III that has robust anti-silencing activity. Deletion mapping revealed that only 0.381 kb in the 5′ portion of the gene (excluding the promoter) is sufficient for barrier activity and that this property is orientation-independent (henceforth referred to as TEL-B). The barrier acivity of TEL-B depends strongly on Sas2p and Esa1p but not on Gcn5p and Sas3p, and is independent of cohesin. Previous investigations have shown that acetylation of H4K16 by Sas2p at subtelomeric regions of chromosome VI leads to deposition of HTZ1 in the nucleosome and its subsequent acetylation by Esa1p of NuA4. All these events together are required to contain the onslaught of telomeric core heterochromatin on neighbouring active regions. Since barrier activity of TEL-B depends on Sas2p and Esa1p, it is possible that TEL-B has the potential to act as a bona fide barrier in situ in its endogenous context. Our hypothesis is further cemented by the observation that there is a physical association between Sas2p, the molecule at the top of the entire cascade of events, with TEL-B by yeast one hybrid analysis. Further experiments will shed light on the role of this sequence in its natural location. In summary, I have identified and characterized two different barrier sequences in S. cerevisiae. Not many barriers are known in budding yeast and there is extensive ongoing research dedicated to understand the mechanism(s) of barrier function. In chapter I of my thesis I present a review of current literature regarding silencing barriers in yeast and other systems. In chapter II I have outlined a detailed characterization of a tDNA barrier element, tRNAGln, present near the silenced rDNA array on chromosome XII. My work addresses the various models for barrier activity and their applicability to the tRNAGln barrier. I have also attempted to understand the role of this tDNA in its natural location on the chromosome with respect to limitation of RDN1 silencing. In chapter III I have described an intensive study of a RNA polymerase II transcribed gene, AAD3, present near the right telomere of chromosome III, which acts as a robust barrier to silencing. I have attempted to answer which mechanism(s) is/are operational at this sequence so as to endow it with barrier potential. My studies with the two barrier elements highlight novel trans-acting factors required for barrier function, differential and selective requirements of certain factors for different barriers, and provide a mechanistic view of the boundary activity of these sequences.
207

Análise estratigráfica do intervalo siluro-devoniano da bacia do Amazonas / Stratigraphic analysis of silurian-devonian interval of the Amazon basin

Werlem Holanda dos Santos 02 April 2014 (has links)
Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / O trabalho consiste na análise estratigráfica do intervalo siluro-devoniano da Bacia do Amazonas utilizando como base os conceitos da Estratigrafia Moderna, mais especificamente a sequência estratigráfica genética, proposta por Galloway (1989), a qual utiliza as superfícies de inundação marinha como os limites de uma sequência sedimentar. A principal razão para a utilização desta metodologia deve-se ao fato que o conteúdo rochoso compreendido no intervalo estudado teve a sua sedimentação relacionada às transgressões marinhas que faziam parte do contexto paleogeográfico da bacia durante o Siluriano e Devoniano. Desta forma, as superfícies de inundação máxima, representativas de eventos cronoestratigráficos, destacam-se nos perfis de raios gama e são tomadas como datum de correlação em treze poços exploratórios, os quais foram distribuídos em quatro seções (A-A, B-B, C-C e D-D) pela bacia. A análise destas seções permitiu a identificação de quatro sequências de terceira ordem (AB, BC, CD e DE), limitadas no topo e na base por superfícies de inundação marinha. Cada sequência é constituída por ciclos regressivo-transgressivos assimétricos, representados pelo trato de sistemas de mar alto e pelo trato de sistemas transgressivo. A análise destas seções integrada à interpretação de mapas estratigráficos (isópacas, isólitas e porcentagem de areias) possibilitou identificar o depocentro da bacia, bem como duas áreas principais como fonte de sedimentos arenosos (uma a oeste e outra a sul). Além disto, foi possível inferir que a comunicação marinha com o continente, durante as transgressões paleozoicas, responsável pela deposição de sedimentos pelíticos, seguiu uma orientação de norte para sul, evoluindo obliquamente em direção ao continente num trend nordeste para sudoeste. Por fim, a partir da análise cíclica em perfis de raios gama, as superfícies de inundação marinha, do intervalo Devoniano, das bacias do Amazonas e Parnaíba foram correlacionadas.
208

Análise estratigráfica do intervalo siluro-devoniano da bacia do Amazonas / Stratigraphic analysis of silurian-devonian interval of the Amazon basin

Werlem Holanda dos Santos 02 April 2014 (has links)
Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / O trabalho consiste na análise estratigráfica do intervalo siluro-devoniano da Bacia do Amazonas utilizando como base os conceitos da Estratigrafia Moderna, mais especificamente a sequência estratigráfica genética, proposta por Galloway (1989), a qual utiliza as superfícies de inundação marinha como os limites de uma sequência sedimentar. A principal razão para a utilização desta metodologia deve-se ao fato que o conteúdo rochoso compreendido no intervalo estudado teve a sua sedimentação relacionada às transgressões marinhas que faziam parte do contexto paleogeográfico da bacia durante o Siluriano e Devoniano. Desta forma, as superfícies de inundação máxima, representativas de eventos cronoestratigráficos, destacam-se nos perfis de raios gama e são tomadas como datum de correlação em treze poços exploratórios, os quais foram distribuídos em quatro seções (A-A, B-B, C-C e D-D) pela bacia. A análise destas seções permitiu a identificação de quatro sequências de terceira ordem (AB, BC, CD e DE), limitadas no topo e na base por superfícies de inundação marinha. Cada sequência é constituída por ciclos regressivo-transgressivos assimétricos, representados pelo trato de sistemas de mar alto e pelo trato de sistemas transgressivo. A análise destas seções integrada à interpretação de mapas estratigráficos (isópacas, isólitas e porcentagem de areias) possibilitou identificar o depocentro da bacia, bem como duas áreas principais como fonte de sedimentos arenosos (uma a oeste e outra a sul). Além disto, foi possível inferir que a comunicação marinha com o continente, durante as transgressões paleozoicas, responsável pela deposição de sedimentos pelíticos, seguiu uma orientação de norte para sul, evoluindo obliquamente em direção ao continente num trend nordeste para sudoeste. Por fim, a partir da análise cíclica em perfis de raios gama, as superfícies de inundação marinha, do intervalo Devoniano, das bacias do Amazonas e Parnaíba foram correlacionadas.
209

Brain Activation Sequences / Brain Activation Sequences

Šusta, Marek January 2017 (has links)
Brain Activation Sequences Abstract INTRODUCTION: This research goes beyond the EEG source localization up to the field of brain connectivity in an attempt to create software tool that eases diagnostic procedures in selected nosologic units by discriminating between patients and healthy controls. METHODS: Experiment 1 - a group of 26 adult patients (14 male, 12 female) suffering from NC and 10 adult controls (5 male, 5 female) participated in the experiment. The experiment contained audio recordings designed to trigger laughter in participants during the EEG recording. Experiment 2 - twenty eight female inpatients diagnosed with ED and ten healthy controls were selected and presented with various stimuli while the EEG was recorded. The Brain Activation Sequences method, applied to all recordings, utilizes nonlinear differential model structure to calculate final output sequence of the brain locations involved substantially in the stimulus processing. RESULTS: Experiment 1 - the BAS results show statistically significant differences in activity between patients and controls namely in gyrus orbitalis, rectus, occipitalis inferior (right), occipitalis medius (right), paracentralis, cinguli, cuneus (right) and parahippocampalis (left). Experiment 2 - the results confirm significant differences in processing the...
210

Subnormality and Moment Sequences

Hota, Tapan Kumar January 2012 (has links) (PDF)
In this report we survey some recent developments of relationship between Hausdorff moment sequences and subnormality of an unilateral weighted shift operator. Although discrete convolution of two Haudorff moment sequences may not be a Hausdorff moment sequence, but Hausdorff convolution of two moment sequences is always a moment sequence. Observing from the Berg and Dur´an result that the multiplication operator on Is subnormal, we discuss further work on the subnormality of the multiplication operator on a reproducing kernel Hilbert space, whose kernel is a point-wise product of two diagonal positive kernels. The relationship between infinitely divisible matrices and moment sequence is discussed and some open problems are listed.

Page generated in 0.4402 seconds